Sponsored by ADI Leak Detection

Why my central heating keeps losing pressure

A sealed heating system should hold its pressure and need topping up once a year at most. This calculator classifies how fast yours is falling and how often you are re-pressurising it, and it deliberately gives you no litres figure from the gauge reading, because a pressure drop cannot be turned into a volume without three things no gauge on a wall can tell you. The volume it does give comes from counting top-ups, and it is small: even a system needing attention every week loses well under 200 litres a year. The reason to act is what that water is doing on its way out, not what it costs.

Last reviewed 14 August 2026

What this calculator tells you

How fast your system is losing pressure, how serious that is, and what it is costing you in water. It does not tell you how many litres the pressure drop represents, and that is the most important thing on this page.

  • The rate of pressure loss in bar per day, when the change is large enough to read
  • A severity classification, taken from the worse of two signals rather than from the gauge alone
  • What that pattern usually indicates and what to check
  • An annual volume and cost, derived from how often you top up rather than from the pressure

Two of the six standard attributes are missing here, and they are missing on purpose. There is no rate of loss in litres per hour and no daily volume, because a gauge reading cannot produce either without three quantities you do not have. Every other calculator on this site reports all six. This one reports four and says why.

How to take the measurement

You need two gauge readings with a known gap between them, and the gap matters more than most people expect.

  • Find the pressure gauge. It is usually on the front of the boiler or on the filling loop pipework near it, and on a sealed system it reads somewhere between about 1 and 2 bar when cold.
  • Take the first reading cold, meaning the heating has been off for a few hours and the pipes are at room temperature.
  • Note the date. The rate depends on the interval as much as on the change, and a remembered "a while ago" makes the answer meaningless.
  • Take the second reading cold as well, at the same sort of time of day, and read the dial square on rather than at an angle.

Both readings cold is the single rule that matters most. A hot system reads higher than a cold one, and on most systems that difference is larger than the change a real slow loss produces over several days. Mix one hot reading with one cold reading and you have not measured a small loss, you have measured the temperature.

How long to leave between them follows from what the gauge can see. A dial marked in steps of about a tenth of a bar cannot show you anything smaller, so the interval has to be long enough for the loss to add up to at least one step. A day is almost never long enough. A week usually is, and for a system that only needs topping up once or twice a year, nothing short of a month will show anything at all.

Getting the cold reading right, and reading a dial without introducing more error than you are measuring, is the whole of the input here. how to read the system pressure gauge covers it properly.

Using the calculator

Most systems are filled to somewhere between 1.0 and 1.5 bar when cold.

Read it cold, with the heating off and the system settled. A hot system reads higher, and comparing a hot reading with a cold one shows a change that is not a leak.

Over the last year or so. This counts for as much as the gauge reading, because it is something you have watched happen rather than a dial read twice.

The figure here changes the annual total more than anything else on this form, so it is worth a moment. If you fill from a bottle or a measured jug, use your own measurement. That figure will be better than any derivation.

Direct evidence that water is leaving the system rather than an inference from the gauge. Ticking it raises the severity by one band, and the result says that this is a judgement rather than a measurement.

It is on your bill. The rate per cubic metre comes from this company's charges scheme.

Rate of pressure loss

Rate of pressure loss is not available.Rate of pressure loss is not available.

Fill in both pressure readings, how long apart they were taken, and how often the system needs topping up. This tool classifies how fast the pressure is falling. It does not convert that into litres, because a gauge cannot support that figure.

What your result means

There are three shapes of answer, and only one of them is a rate. The other two are worth as much, because each rules something out.

ResultWhat it means
A measurable lossThe pressure fell by more than the gauge step over the interval, so there is a rate to report and a band to put it in.
Below what the gauge can resolveThe change is smaller than one step on the dial. That is not a small loss, it is no measurement. Leave it longer and read again, cold both times.
The pressure has risenIt reads higher now than at the last top-up. Usually a hot reading compared against a cold one. If both were cold, look at the filling loop and the expansion vessel charge.
The three results and what each one is

The severity is not taken from the gauge alone. It is the worse of two signals: how fast the pressure is falling, and how often you have to top the system up. Visible damp or staining anywhere near pipework raises it one band, which is a stated judgement rather than a measurement, and the result says so.

Top-up frequency carries at least as much weight as the pressure reading, and that is deliberate rather than a compromise. A frequency is a behaviour you have watched over months. A pressure rate is a coarse dial read by eye twice. The input that looks vaguer is the better evidence, because of what each one is a measurement of.

A system that only loses pressure while the heating is running is a separate case with a different answer, and it is common enough to be worth checking before anything else. why pressure drops only when the heating is on sets out what that pattern points at.

How the calculation works

Two separate calculations that never touch each other. One classifies the pressure, one counts the water, and no number crosses between them.

  • The rate = the fall in pressure ÷ the days between the two readings, in bar per day
  • The band = which of four rate ranges that falls in, then taken as the worse of that and the top-up frequency
  • The annual volume = top-ups per year × litres per top-up
  • The cost = that volume in cubic metres × your volumetric rate

Why this calculator does not give you a litres figure

Because it cannot be done honestly, and every tool that appears to do it has invented the missing parts.

To convert a pressure drop into a volume of water you need three quantities, and a gauge on the wall gives you none of them.

What you would needWhy you do not have it
The volume of water in the systemIt depends on the length and bore of every pipe, the type and size of every radiator and the boiler itself. Nobody has this figure for their own house, and estimates of it vary by a factor of several.
The air charge in the expansion vesselThe vessel absorbs the expansion of the water and is what makes the pressure move at all. Its charge is not marked on the gauge, changes over the life of the system, and is the thing that has often failed in the first place.
The water temperature at both readingsPressure in a sealed system moves with temperature independently of any water being lost. Without both temperatures you cannot separate the two.
What a litres figure would require

A calculator that hands you litres from a bar reading has quietly assumed all three. It has picked a system volume, picked a vessel charge, and assumed both your readings were at the same temperature. None of those assumptions is shown to you, and the answer moves enormously depending on what they were.

So the annual volume on this page comes from somewhere else entirely: how often you top the system up, multiplied by how much water goes in each time. Both are things you can actually observe, the second is an adjustable assumption you can change, and neither is derived from the pressure reading. That figure is honest about being an estimate. A litres figure from the gauge would not be an estimate, it would be a guess dressed as a measurement.

It works as a measure of the loss because a sealed system has no legitimate consumption. Nothing draws water off it for use. So over a year in which the pressure ends roughly where it started, every litre you put in replaced a litre that left, which is exactly the quantity you wanted.

Assumptions and sources

Per-event central heating top-up volume, derived

Derived by Q3 from Boyle's law and typical UK domestic expansion vessel specifications · derived source

  • Water held in an expansion vessel: V_water = V * (1 - P0 / P), pressures absolute equation
  • Typical top-up, 8 l vessel at 1.0 bar pre-charge, 0.5 to 1.5 bar: 1.6 litres
  • Typical top-up, 8 l vessel at 0.75 bar pre-charge, 0.5 to 1.5 bar: 2.4 litres
  • Typical top-up, 12 l vessel at 1.0 bar pre-charge, 0.8 to 1.5 bar: 2.4 litres
  • Typical top-up, 8 l vessel at 1.0 bar pre-charge, 1.2 to 1.5 bar: 0.9 litres
  • Realistic range: 0.9 to 2.5 litres

Litres in a cubic metre

Exact by definition

SI definition. The litre is defined as one cubic decimetre, so one cubic metre is exactly 1,000 litres.

Pascals in a bar

Exact by definition

Definition of the bar as exactly 100,000 pascals, that is 100 kilopascals.

Days used for an annual figure

Editorial convention

Annual figures on this site are a daily rate multiplied by 365. Leap years are not modelled, so a leap year is understated by one day, which is 0.27 per cent.

Smallest pressure change you can read

Illustrative assumption, editable above

Illustrative assumption. A domestic filling loop gauge is a dial marked in steps of about 0.1 bar or coarser, and it is read by eye at an angle. A change smaller than one step on the dial cannot be distinguished from a reading error, so this calculator declines to report a rate below it.

Highest system pressure the form accepts

Editorial convention

A typing check rather than a physical limit. A sealed domestic heating system gauge reads well below this, so a larger number is almost always a decimal point in the wrong place or a reading in a different unit.

Top-up events per year by frequency

Editorial convention

Each frequency option is converted to events per year using the midpoint of the interval it describes, so that the annual top-up volume is derived from how often you top up rather than from any pressure reading.

Two of the figures here are ours rather than anybody else’s, and they are the two that most affect the answer, so they are named plainly.

  • The water added per top-up, taken as 1.5 litres. No published measurement of this was found. It is derived from the gas law that governs an expansion vessel, for a typical 8 litre vessel pre-charged to 1 bar and topped up from 0.5 to 1.5 bar, and across the usual vessel sizes and pre-charges the realistic range is about 0.9 to 2.5 litres. It ships as an adjustable input for that reason.
  • The four pressure loss bands. These are this project’s own thresholds and no published classification has been found to cite for them. They are stated as a working scale rather than as a standard, and if a sourced one turns up they will be replaced by it.

The refusal above is enforced rather than promised, and it can be checked. The one function on this site that returns a volume for a heating system has no pressure field on its input type at all, so it is structurally incapable of deriving a volume from a pressure. A test moves both pressure readings across their entire range and asserts that the litres figure does not change. An assurance in prose would decay quietly. That will not.

The gauge resolution of a tenth of a bar is also an assumption rather than a specification. It describes a typical domestic dial marked in steps of about that size and read by eye. If your gauge is digital, or finer, it can resolve a smaller change than this calculator will report.

What to do next

Four things, and the first two cost nothing and rule out the two commonest causes.

  • Look at the outside end of the pressure relief discharge pipe, which terminates where you can see it, usually on an external wall. Water or staining there means the system has been discharging through the valve, which is water genuinely leaving with nothing wrong with any pipe.
  • Take two cold readings a week apart and write both down with their dates. Most people compare a hot reading with a cold one and measure the temperature instead of the loss.
  • Count how often you have topped it up over the last year. That number carries more weight than the gauge and it is the one you already have.
  • Look for damp, staining or a tide mark near pipework, under floors and in airing cupboards. Water leaving a heating circuit is hot, so it evaporates and leaves a mark where a cold leak would leave a puddle.

If none of that finds it, the loss is somewhere you cannot see, and there is no household method that locates a buried circuit. how a loss from a heating circuit is traced explains what the available methods can and cannot establish.

One case does reach your water meter, and it is worth knowing which. a system topped up automatically shows the loss at the meter, because an automatic filling device replaces the water continuously instead of in one lump when you open a filling loop by hand.

Common questions

How much water am I actually losing?
Not something a pressure gauge can tell you. Turning a pressure drop into litres needs the volume of water in the system, the air charge in the expansion vessel and the water temperature at both readings, and none of those can be read off the dial. What this calculator does instead is count something you can observe: how often you top the system up, and roughly how much water goes in when you do.
How often should a sealed system need topping up?
Ideally never, and about once a year at the outside. A sealed system has no legitimate consumption, so unlike a household supply there is no normal use to subtract. Every litre you put in replaced a litre that left. Anything at once a month or more is a system losing water somewhere and worth investigating.
My gauge has not moved in three days. Does that mean there is no leak?
No, and this is the most misread result on the page. A typical dial is marked in steps of about a tenth of a bar, so a change smaller than that cannot be distinguished from a reading error. A system that needs topping up monthly is falling by roughly a thirtieth of a bar a day, which takes about three days to add up to one visible step. If you top up regularly and the gauge looks still, you almost certainly read it too soon rather than having nothing wrong.
The pressure is higher than when I topped it up. What does that mean?
Most often that one of the two readings was taken while the water was hot. Water expands as it heats and a sealed system has nowhere to send that expansion except the expansion vessel, so the gauge reads higher hot than cold. Compare two cold readings before reading anything into it. If both were genuinely cold, a filling loop left open and an expansion vessel that has lost its air charge are the two usual causes.
Is a heating leak expensive in water?
No, and this site would rather say so. At the assumed volume per top-up, a system needing attention every week loses about 78 litres a year, and one needing it more than weekly about 156. A tap dripping four times a minute wastes more than three times that. The cost of a heating system losing pressure is the damage the water does where it escapes, not the water itself.
Does water lost from the heating show up on my water meter?
Only if the system refills itself. A sealed circuit is filled by hand through a filling loop that is normally closed, so the water crosses the meter in one lump when you open it and is invisible the rest of the time. A system fitted with an automatic filling device tops itself up without telling you, and that loss does pass the meter continuously, which is the one case where an overnight meter test will find it.