Water takes the cheap route
A heating circuit does not share water out fairly. It shares it out by resistance, and resistance is a property of geometry: the first radiator off the flow, on a short tail, in fat pipe, gets far more than its share. The one at the end of forty metres of microbore with six bends in it gets whatever is left after everything upstream has helped itself.
The symptom set is always the same and it is always misread. Near rooms overheat, so their thermostatic heads get wound down to two. Far rooms never reach setpoint, so somebody raises the boiler flow temperature. Everything now runs hotter, the near rooms overheat again at a lower valve setting, standing losses go up, and a condensing boiler stops condensing. The system has been repaired into a worse condition than it started in.
Balancing reverses that by deliberately adding resistance to the circuits that have too little, until each emitter carries the flow it was sized to carry. Nothing is being made more efficient in the process. Flow is being taken away from the rooms that were stealing it and given to the rooms that were paying for it.
Two things are commonly offered as substitutes and neither one is. Thermostatic radiator valves are a room control, and a modulating one at that; they close down a hot room only after it has already been overheated, and they do it against a differential pressure the balance is supposed to have set.
A variable-speed circulator changes how much total flow arrives at the manifold; it has no opinion whatsoever about how that flow divides once it is there.
Nobody can see flow
The central difficulty of this work is that flow rate is invisible and every instrument for it is indirect. On a commercial system with fixed-orifice commissioning valves, a differential pressure set across the valve tappings and the manufacturer's chart give a genuine flow reading. On a domestic system fitted with plain lockshields, no such tapping exists, and you are working from temperature.
Temperature difference is the usable proxy. An emitter carrying its design flow, fed at its design flow temperature, will drop the water by the design amount across itself — commonly in the region of a ten to twenty degree Celsius drop on a boiler system, and much less on a heat pump system. Too much flow and the drop shrinks toward nothing, with the return almost as hot as the flow. Too little and the drop grows, the tail end goes cold, and the bottom of the emitter is where you feel it.
Measure it with two matched surface probes, clamped on clean pipe under the insulation, at the same distance either side of the emitter, and give each one time to settle. A pair of infrared readings taken across painted pipe at two different angles is not a measurement of anything. Where the system is old enough to have valves with no tappings and pipes that will not take a clamp, a portable ultrasonic meter on a straight run is the honest fallback. Whatever instrument you use, the reading is only meaningful against a stable supply. Balancing while the boiler is cycling on its own thermostat, or while a weather compensator is dropping the flow temperature through the afternoon, produces a set of numbers that describe the weather. Lock the plant at a fixed flow temperature for the duration.
What each emitter is owed
Design flow per emitter is not a preference; it falls out of arithmetic that has already been done. The emitter has a rated output at a stated mean water temperature. The system has a design temperature drop. Flow rate is the output divided by the product of the water's specific heat capacity and that drop. Every balancing valve setting on the job exists to deliver one of those numbers.
Two corrections trip people up. The first is that a radiator's catalogue output is quoted at a mean water temperature well above what many systems now run, and output falls steeply as that mean drops — a radiator sized for a seventy-five degree flow gives a fraction of its rating at fifty. The second is that output and flow do not scale together: halving the flow through an emitter does not halve its output, it reduces it by rather less, which is exactly why a system can be badly out of balance and still, more or less, heat the house.
Low-temperature systems change the arithmetic in a way that catches out anyone carrying boiler habits into a heat pump retrofit. Holding output while cutting the temperature drop from twenty degrees to five means roughly four times the flow through the same emitter, through the same pipe, with roughly thirteen times the friction loss, because pressure loss climbs with flow to a power near 1.85. Circuits that were merely unbalanced on a boiler become genuinely undeliverable on a heat pump, and the fix is pipework, not a valve setting.
Write the design flow for every emitter on a schedule before touching a single valve. Balancing without target figures is not balancing; it is making the radiators feel similar to the back of a hand, which is a test that fails at about a twenty percent error and passes everything worse than it by accident.
The flow each emitter is owed, one line of the schedule at a time: its output at the system's real mean water temperature and the design drop give the flow, and the bore feeding it gives the velocity — four times the water at a heat pump's 5 K (9 °F) as at a boiler's 20 K (36 °F).
The emitter's output at the design water temperature, or the whole circuit's load.
The difference between flow and return water temperatures the system is designed for.
The inside diameter of the pipe carrying this flow, not its nominal size.
Design flow rate
1.893 gal/min
Flow = heat ÷ (the specific heat of water × the temperature drop). Halve the drop and the flow doubles; a quarter of it, as a heat pump runs, is four times the water through the same emitter and the same pipe.
- Water velocity in the pipe
- 3.09 ft/s
- Mass of water per second
- 0.26 lb/s
They open the calculator with your figures already in it
Heating Circuit Flow Rate Calculator (Heat Output, ΔT and Pipe Velocity): 1.89 gal/min — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 1.893 gal/min — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- Water is taken at a kilogram a litre — 8.34 lb a US gallon. Hot water is slightly lighter — about 2 per cent at 70 °C (158 °F) — so the volume flow at working temperature is that much higher than shown.
- Glycol mixtures carry less heat per degree than water and are more viscous; a system with antifreeze needs the fluid maker's specific heat and a larger flow.
- The velocity is reported, not judged. Noise, erosion and the pump's duty set the limits for a given pipe, and the pressure loss that comes with the velocity is a separate calculation.
The circuit that sets the pump
One circuit in the system has the greatest total resistance at design flow, and that circuit — the index — determines the head the circulator must produce. Everything else in the system has surplus pressure available to it, and balancing is the business of absorbing that surplus deliberately at a valve instead of letting it be absorbed accidentally by excess flow.
The index circuit is frequently not the longest one. Resistance is length plus fittings plus terminal units plus whatever the water passes through at the plant, and a short run through undersized pipe with a heat exchanger and four elbows in it will beat a long run in generous pipe. Microbore tails, a plate heat exchanger, a zone valve and a strainer all put resistance in places a tape measure cannot see.
Count fittings off the route as built. Elbows, tees taken on the branch, isolating valves either side of every emitter, the lockshield itself, the entry and exit of the emitter — each converts to an equivalent length of straight pipe, and on a domestic circuit those equivalents routinely exceed the straight pipe they are added to. A drawn route rarely survives a joist run intact, so walk the pipe.
Identify the index by measurement where you can, not by inspection. With every regulating valve fully open and the system stable, the circuit with the lowest ratio of measured flow to design flow is the index, and no other valve in the system will ever be opened further than that one.
Bends, tees and the isolating valves either side of every emitter carry more of the index circuit's resistance than the straight pipe does, so count them off the as-built route.
The total count of 90° elbow fittings in the piping run.
The straight-pipe length that produces the same friction loss as one elbow.
The total count of tee fittings where flow branches off the run.
The straight-pipe length that produces the same friction loss as one branch-flow tee.
Total fitting equivalent length
12 ft
They open the calculator with your figures already in it
Pipe Fitting Equivalent Length Calculator: 12 ft — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
What this calculation does not cover
- Totals fittings, not the components that usually dominate the head. A coil, heat exchanger, strainer, balancing valve or control valve is published as a pressure drop at a stated flow - feet of head or kPa, not an equivalent length - and on a typical hydronic circuit those together exceed the whole pipe-and-fitting figure. A pump chosen from this total plus the straight pipe alone lands short, and the shortfall turns up as the branch that will not balance.
- One path, not a system. Pump head is set by the INDEX circuit, the single worst route from the pump out to the furthest or most restrictive terminal and back, so the fittings that belong in this box are the ones along that path only. Adding up every elbow in the building inflates the total enormously and buys a pump that overpumps every other branch; the parallel branches get balanced down to the index circuit, they do not add to it.
Head for the system you built, not the one drawn
A circulator meets a system at one point: where its pump curve crosses the system's resistance curve. That crossing is a physical fact about the installation, and it moves whenever the installation differs from the drawing — an extra pair of bends to clear a steel, a size dropped on a leg because that was what the van had, a filter added at commissioning that nobody costed for pressure.
This is why the friction loss rate has to come from the pipe actually installed at the flow actually required, and why the equivalent length has to come from the route actually taken. Assemble those two and the duty is a number you have derived. Read it off a rule of thumb instead and you get a circulator that is either loud or short, and no way to tell which until the readings come in.
The common failure is an oversized pump treated as insurance. Excess head does not distribute itself evenly; it goes into the circuits with the least resistance, driving flow well past design in the near emitters and producing velocity noise in the pipe — that persistent rush behind the skirting that nobody can locate. Fitting a larger pump because a far radiator is cold is the exact opposite of the correct move: the far radiator is cold because the near ones are taking too much, and more head gives them more still.
Where the circulator is a modern variable-speed unit, its control mode is part of the balance. Constant-pressure mode holds a fixed differential no matter how many zones are closed; proportional-pressure mode drops the head as flow falls, which suits a system with thermostatic valves closing all over it. Set the mode and the curve to give the index circuit its design flow, and let the balance hold everything else in proportion.
The duty you need is the index circuit's resistance at design flow — a number the installation produces, not one the catalogue supplies.
The friction head loss per 100 ft (or 100 m) of pipe run, from a pipe sizing chart or a Hazen-Williams friction loss calculation at the design flow rate.
The straight pipe length plus the equivalent length of all fittings, valves, and equipment in the circuit's index (longest/most-resistant) run.
Total pump head required
10 head units
Friction loss rate must come from your system's actual pipe sizing chart or Hazen-Williams calculation for the design flow rate — this calculator only scales that rate by total equivalent length; it does not calculate the friction loss rate itself.
They open the calculator with your figures already in it
Hydronic Circulator Pump Head Loss Calculator: 10 head units — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Estimated cost — your price
This site holds no price list for this material — local prices vary too much to publish honestly. Enter your supplier's price and the result is costed with it.
What this calculation does not cover
- Head on its own selects no circulator. A pump is chosen where the required head and the design flow meet on its curve, so this figure has to be carried forward with the same flow rate the friction loss rate was read at — two systems needing identical head at very different flows take entirely different pumps.
- Plain water is assumed. A 30 to 50 percent propylene glycol charge is thicker and heavier, raising circuit friction at the same flow and cutting the flow any given pump delivers, so a snowmelt slab or a freeze-protected loop sized from a water friction rate ends up short of flow at the far end of the circuit.
- In a closed loop the pump lifts nothing — the down-leg balances the up-leg — so building height belongs nowhere in this figure and adding it oversizes the pump. The reverse applies to an open circuit such as a drainback array or an open tank: there the vertical rise from the water surface to the discharge is real head, and it is not included here.
Valves that make the flow adjustable
A balance is only as good as the device holding it. A plain lockshield with a shroud over it is adjustable, repeatable to about a quarter turn if you are careful, and carries no way of knowing what flow that quarter turn corresponds to. It is what most domestic systems have, and it works when paired with a temperature measurement and a patient method.
A double-regulating valve improves on that by separating the setting from the isolation: you can shut it to work on the emitter and reopen it to exactly the setting it held, which removes the single most common way a balance gets destroyed during a later repair. A fixed-orifice commissioning valve adds pressure tappings and a published flow chart, turning the setting into a measured quantity instead of an inferred one. Pressure-independent control valves take a different route entirely, holding a set flow regardless of what the differential across them does. On a system with many zones opening and closing they remove most of the interaction that makes proportional balancing tedious, and they cost more per terminal and need a minimum differential to work at all. They are a design decision, not a remedy to apply to a system already installed.
Thermostatic radiator valves belong on the flow side, and their behaviour depends on valve authority: the share of the circuit's pressure drop that the valve itself accounts for when open. Give a thermostatic head too little authority and it goes from fully open to fully shut across a couple of degrees, hunting the room around setpoint. That authority is set by the balance, which makes the balance a comfort issue as much as an efficiency one.
The proportional method, in the order it works
Proportional balancing exists because circuits interact. Throttle one and the pressure available to its neighbours rises, so their flows rise too, and a method that sets each valve once and moves on will never converge. The proportional method accepts the interaction and works with it: every terminal on a branch is brought to the same ratio of measured to design flow, and that ratio is then corrected for the whole branch at once with a single upstream adjustment.
The consequence is that the index valve is never throttled. It stays fully open, sets the reference ratio, and the whole system is scaled to it at the pump at the end. Any method that ends with the index circuit partly closed has thrown away head that the pump is still generating and the electricity bill is still paying for.
Expect the last pass to be small. The first pass moves readings by large fractions, the second by a few percent, and if the third pass is still moving things substantially, the problem is not the balance — it is air, a partly blocked strainer, a valve that will not hold, or a pump control mode that changes the differential every time you adjust something.
On a small domestic system with lockshields and surface probes the same logic applies with temperature standing in for flow: bring every emitter to the same temperature drop across itself, working from the far end back, then set the flow temperature and pump so the index emitter delivers its output.
- Fill, treat, vent and bring the system to a fixed flow temperature with every emitter calling and every regulating valve fully open.
- Measure the flow, or the temperature drop, at each terminal and record it against that terminal's design figure as a ratio.
- Identify the index: the terminal with the lowest ratio while fully open. Its valve stays open for the rest of the exercise.
- Working from the far end of each branch back toward its connection, regulate each terminal until every one on that branch carries the index ratio.
- Balance branch against branch, then riser against riser, using the same ratio logic, finishing at the circulator.
- Set the pump curve or speed so the index terminal receives its design flow, and confirm the rest have followed in proportion.
- Lock every valve, mark the setting on the body or on a label, and enter it on the commissioning sheet.
Differential pressure and the zones that close
A balance set with everything open describes a condition the system rarely occupies. Thermostatic heads shut as rooms reach setpoint, zone valves close on their own schedules, and each closure takes flow out of the system and pushes the circulator back up its curve. Differential pressure across the remaining circuits rises, the flows through them rise, and the balance drifts away from the one that was commissioned.
The audible consequence arrives first. Thermostatic valves nearly closed against a high differential whistle, and the noise gets blamed on the valve. The thermal consequence follows: the rooms still calling get more flow than they need at exactly the moment the plant is running at reduced load, which on a condensing boiler raises the return temperature and stops it condensing.
Differential pressure control valves answer this by holding a fixed differential across a branch no matter what happens downstream, which keeps the balance within that branch intact across the whole range of zone positions.
On a small domestic system the same job is attempted by an automatic bypass valve, and it is worth being clear about what that device does: it protects the pump and the boiler by dumping flow straight from the flow to the return when everything closes. Set too low, it dumps hot water back to the boiler continuously and ruins both the balance and the return temperature.
A variable-speed circulator on proportional-pressure control does most of this for nothing on a domestic system, provided the curve was set from the index circuit rather than left on whatever the factory shipped. Check the pump's own display for its operating point at part load; on most modern units that reading is free and nobody looks at it.
Air is a fault, not a nuisance
Air in a heating circuit is not an inconvenience to be bled out on a Saturday. It is a blockage that moves. Gas collects at the high points and at the tops of emitters, occupies volume that water should have, and dramatically changes the resistance of whichever circuit it is currently sitting in — which is why a system that will not hold a balance from one visit to the next is so often a system that keeps taking air on board.
Where the air comes from is the more useful question. Water released from solution as it heats accounts for the first few weeks after filling and then stops. Air still arriving after that has a source: a leak on the negative-pressure side of the pump drawing air in without leaking water out, an expansion vessel that has lost its charge and is letting the system go into vacuum when it cools, or a top-up habit that introduces fresh oxygenated water every few weeks.
Deal with it at the plant, not at the radiators. A micro-bubble separator placed where the water is hottest and the pressure lowest — typically on the flow immediately after the heat source — strips bubbles continuously and takes them out of circulation. Automatic vents at genuine high points handle what collects there. Manual bleeding is a symptom check, not a cure.
Dirt does the same damage more slowly. Magnetite generated by corrosion settles in the bottom of emitters and in the crevices of valves, changing resistances one circuit at a time until a balance set two winters ago describes nothing. A magnetic filter, a proper flush and a correctly dosed inhibitor to a recognised code of practice keep the balance you set worth having.
A separator only strips bubbles at the velocity it was sized for; match the connection size to the system's design flow and its position in the circuit follows from that.
The hydronic system's design flow rate through the separator.
The maximum velocity through the separator's connection for effective air and dirt removal — commonly around 0.3 m/s (about 1 ft/s).
Minimum connection diameter
5.73 in
Effective air/dirt separation requires low velocity at the separator itself — confirm the manufacturer's rated maximum velocity and connection size for your specific separator model rather than assuming a generic value.
They open the calculator with your figures already in it
Hydronic Air/Dirt Separator Sizing Calculator: 5.73 in — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 5.73 in — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- Position beats size here. Air leaves solution where the water is hottest and the pressure lowest — on most systems the boiler or chiller outlet, upstream of the circulator. Put the same separator on the cool return and it sees air still dissolved and passes it straight through. No connection diameter compensates for the wrong location; the low velocity sized here only earns its keep once the bubbles actually exist.
- This is a connection diameter, not a pressure drop. Every separator adds head to the circuit — the vessel, the coalescing medium inside it, and the expansion and contraction at each end — and none of that comes back from a velocity formula. A pump selected on pipe friction alone can land short of its duty point once the separator, the strainer and the dirt the separator will eventually be holding are all counted in.
- The dirt half is a different problem from the air half. What blocks a modern condensing boiler's heat exchanger is magnetite — fine ferrous sludge that a gravity pocket catches poorly and a magnet catches well — and no velocity calculation tells you whether you need the magnetic version. Nor does it size the dirt pocket or set how often it is flushed: a separator that is never blown down is a reservoir, not a filter.
Cold fill, expansion, and the pressure it all rests on
Balance readings that drift all afternoon are frequently a pressure problem wearing a flow problem's clothes. Cold fill pressure has to be high enough that the highest point of the system stays positively pressurised when everything is cold, with margin for the pump's suction side. Set it too low and the top of the system goes into partial vacuum as it cools, drawing air in through vents that were fitted to let air out.
The expansion vessel's pre-charge is checked with the system pressure removed from the vessel, and it should match the cold fill pressure. A vessel that has lost its charge, or one with a perished diaphragm, is a vessel full of water: the system then has nowhere to expand into, pressure climbs as it heats, the relief valve lifts, and the system loses water it will have to be topped up with. Each top-up is a fresh dose of dissolved oxygen and the corrosion that follows.
Vessel size comes from the system's water content, which on a job with cast iron emitters or a buffer is a great deal more than anyone estimates. Add up pipe volume, emitter volume, the heat source and any store; an undersized vessel does its job right up until the coldest week of the year.
Pump position matters to the same argument. A circulator pumping away from the point where the expansion connects raises pressure through the whole distribution; pumping toward it lowers pressure through the distribution and can pull the top of the system below atmospheric while the gauge at the boiler still reads healthy. On a retrofit where the pump has been moved, this is worth checking before any valve gets touched.
Readings that will not settle are often a pressure fault, so size the vessel against the water the system genuinely holds — pipe, emitters, heat source and store together.
The total water volume held in the closed hydronic loop, including boiler/chiller, piping, and terminal units.
The fraction by which the system water expands in volume across its full operating temperature swing.
The fraction of the tank's total volume that is actually usable to accept expanded water, based on the tank's fill and relief pressures.
Minimum expansion tank total (nominal) volume
8.67 gal
This is a simplified sizing approximation — final expansion tank sizing should follow the ASHRAE/hydronic system design method (or the tank manufacturer's sizing software) accounting for actual fill pressure, relief valve setting, and system operating temperature range.
They open the calculator with your figures already in it
Hydronic Closed-Loop Expansion Tank Sizing Calculator: 8.67 gal — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Estimated cost — your price
This site holds no price list for this material — local prices vary too much to publish honestly. Enter your supplier's price and the result is costed with it.
What this calculation does not cover
- The expansion fractions offered here are water. A loop charged with 30 to 50 percent propylene or ethylene glycol expands appreciably more over the same temperature swing, and the fluid supplier publishes the figure for the concentration you are running. Size a glycol system on a water fraction and the relief valve opens at design temperature and dumps the glycol you paid for — the system then makes up with plain water and quietly loses its freeze protection.
- A tank of the right volume does nothing if it is precharged wrong. The air side must be set to the system's fill pressure at the tank's own elevation, with the water side empty, before the loop is filled. A tank shipped at 12 psi (83 kPa) and hung on a system filled to 25 psi (172 kPa) has already spent most of its acceptance volume before the water is even warm, and the acceptance factor above has stopped describing it.
- Where the tank ties into the loop is not a sizing question, but it decides whether the tank works. The connection point is the one place in the system whose pressure the pump cannot change — pipe the tank into the pump's suction side and the pump adds its head to the rest of the loop, put it on the discharge side and the pump subtracts it, which can drop the pressure at the top of the system or at the pump inlet far enough to draw air or cavitate on a tank that was sized correctly.
What a system that will not balance is telling you
There is a point in this work where the correct answer stops being another quarter turn. If every valve on the system is throttled hard and the index circuit is still short, the balance is not the problem. Balancing redistributes flow; it cannot create any. A system that cannot deliver design flow to its index circuit with the index valve wide open has a plant or a pipework fault, and turning valves is now just moving the shortage around.
Read the pattern, not the individual reading. Flow that collapses when one zone closes points at differential control, not at balance. Readings that change between the morning and the afternoon point at air, at a valve that will not hold its setting, or at a compensator quietly changing the flow temperature underneath the measurement. A single emitter cold at the bottom is sludge; cold at the top is air; cold overall with a good drop across it is an emitter too small for the water temperature the plant now supplies.
The makeup water meter is the single most useful and least used diagnostic on a heating system. A system that needs topping up on any regular schedule is either leaking or venting through its relief valve, and both of those explain a balance that will not hold. Meter the top-up, quantify the rate, and the answer to whether you are chasing a leak or a failed expansion vessel usually arrives before the first valve is touched.
Finally, be willing to conclude that the pipework is wrong. A pair of emitters teed off a single undersized leg, a microbore manifold serving a room that has since been extended, a heat pump grafted onto a distribution designed for seventy-five degree water — none of these are balancing problems, and presenting them as such wastes a day and leaves the customer with the same cold room.
| What you observe | What is actually happening | What to measure next |
|---|---|---|
| Every valve throttled hard and the index still short | The circulator is on the wrong curve, or the index leg is undersized pipe | Head at design flow against the pump curve for the unit installed |
| Flow collapses in the rooms still calling when a zone closes | No differential control; the pump rides up its curve into what remains | Differential pressure across the index branch, zones open and then closed |
| Readings drift across a single working day | Air still coming out of solution, or plant changing the flow temperature | Separator position, vent operation, and whether the compensator is locked |
| Emitter cold at the bottom, good flow temperature at the tail | Sludge occupying the bottom of the emitter, not a flow shortage | Return temperature at the tail against the flow temperature at the head |
| Emitter cold at the top only | Air, and air that keeps arriving has a source | Cold fill pressure and expansion vessel pre-charge with the system drained of pressure |
| Balance holds but the rooms never reach setpoint | Emitters sized for a higher mean water temperature than the plant delivers | Emitter output at the actual mean water temperature, room by room |
| The system is topped up every few weeks | A leak, or a failed vessel venting through the relief valve | Makeup volume against the interval, taken before any valve is adjusted |
A makeup rate is the cheapest diagnostic on a heating system: put the top-up volume against the interval and the answer says whether you are hunting a leak or a dead vessel.
The total volume of water held in the boiler/hydronic system's piping, boiler, and terminal units.
The expected fraction of total system water volume lost per period from minor leaks, air venting, and maintenance draindowns.
Estimated makeup water needed
10.6 gal
Loss rate percentage varies widely by system age, condition, and maintenance practice — an unusually high or increasing makeup water demand can indicate a leak that should be investigated rather than simply compensated for.
They open the calculator with your figures already in it
Boiler Makeup Water Calculator: 10.6 gal — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Estimated cost — your price
This site holds no price list for this material — local prices vary too much to publish honestly. Enter your supplier's price and the result is costed with it.
What this calculation does not cover
- The whole calculation is one multiplication of system volume by the loss percentage, so operating pressure, fill-loop and feed water temperature, and the behaviour of the expansion vessel never enter it — a system that cycles pressure hard and one that sits stable return the same volume if you type the same two numbers.
- Loss rate is entered as a percentage "per period" but no period is ever asked for and the answer comes back as a plain volume, so the figure carries whatever window you had in mind — annual, heating-season or monthly — and a figure quoted on one window cannot be compared with a system measured on another.
- The loss-rate box accepts 0.5 to 5 and rewrites anything outside that band to the nearest bound when you leave the field, showing a notice such as "Using 5 — the highest this calculator accepts. You entered 10", so a system genuinely shedding 8 percent has to be modelled by scaling the volume figure instead, and the 0.5 floor means a tight sealed system cannot be expressed as needing nothing.
- System water volume is taken exactly as supplied and is never derived from or checked against pipe runs, boiler content or terminal-unit capacity, so a volume estimated loosely from design drawings passes into the makeup figure one-for-one — understate the system by a fifth and the water is understated by a fifth.
- The result arrives as a single bulk volume with no breakdown behind it, so the share going to air venting, to maintenance draindowns and to actual leakage stays merged in one number even though the loss-rate help text names all three as separate causes, and it will not tell you which of them to chase.
The record, and the valves left locked
A balance that is not written down has a life expectancy of one radiator change. The commissioning record needs the design flow and the achieved flow for every terminal, the valve setting that produced it, the flow temperature the work was done at, the pump mode and curve, and the date. Anything less and the next person has to start from nothing, which in practice means they will not start at all.
Mark the settings physically as well as on paper. A shroud replaced over a lockshield, a numbered indicator on a double-regulating valve, a label on the manifold — whatever the valve type allows. The point is that a plumber replacing an emitter in three years can isolate at the valve and put it back where it was, instead of leaving it wherever it happens to land.
Tell the occupant what the lockshields are and why they are not a control. The thermostatic head is theirs to turn; the shrouded valve at the other end is not. This one sentence at handover prevents a surprising proportion of the systems that come back out of balance. Balance is worth revisiting after the first full heating season, and after any change that alters resistance anywhere: a new emitter, a heat source swap, a zone added, a filter fitted. Each of those moves the system curve, and a system curve that has moved is a set of readings that no longer describes the building.
What has to be on the van before the first valve moves
Balancing is a measuring job with a spanner in the other hand. None of this is a material take-off; it is the set of figures and instruments that make a reading mean something.
- Design flow for every emitter, written on a schedule — Output divided by specific heat times the design temperature drop, corrected for the mean water temperature the plant actually delivers.
- The index circuit's equivalent length, fittings counted — Isolating valves, tees taken on the branch and emitter connections usually outweigh the straight pipe on a domestic leg.
- The pump curve for the circulator that is genuinely fitted — Together with its control mode; proportional and constant pressure behave differently as zones close.
- A differential pressure set matched to the valves on site — Fixed-orifice commissioning valves only give a flow reading against their own published chart.
- Two matched surface probes for emitters with no tappings — Clamped on clean pipe under the insulation, equal distances either side, given time to settle.
- System water volume, for the vessel check and the inhibitor dose — Pipe plus emitters plus heat source plus any buffer; cast iron and buffers wreck estimates made by eye.
- A makeup water reading at the start and again at the end — A rising rate reframes the whole job from a balancing exercise to a leak or a failed expansion vessel.
Opens the calculators above on one screen with the dimensions from this article already filled in. Quantities only — this site publishes no price list, because local prices vary too much to publish honestly.