Compressed Air Leak Test Calculator

Shut the compressor off at the end of a shift, leave every drop and tool alone, and come back in ten minutes. Whatever the gauge has lost went out through holes. That is the whole test, and it works because the system is a known volume: air stored between two pressures is arithmetic, and dividing it by the clock gives a flow rate that no amount of arguing about hearing hisses can produce. The awkward part is that the answer usually comes back bigger than anyone expects, and the hole it corresponds to is smaller than anyone expects.

Receiver, wet tank, dry tank and the piping between them, added up. The receiver capacity is on the vessel; the piping is bore area times length. Piping is small next to a receiver and easy to leave out, and leaving it out understates the leak.
Stopwatch, not an estimate. This divides straight into the answer.
Anything that drew air during the test window and was not a leak — a dryer purging, an automatic drain firing. Leave at zero for a properly isolated test.
Free air volumes are referenced to the air outside. Around 12.2 psia at 5,000 ft, and every free-air figure on the page moves with it.
Leak flow through a small hole tracks absolute pressure, so the rate measured across the test band is scaled to this.
Optional. The delivered figure for your machine at your pressure, from the data sheet or a pump-up test. Displacement is not delivery.
Optional. Motor input power divided by delivered flow, times 100. Measure it with a clamp meter, or take both figures from the data sheet.
Leaks run whenever the system is up, including nights and weekends if nobody shuts the valve.
All-in from the bill, not the headline energy rate.
Only used for the equivalent hole size at the bottom. A ragged split in a hose passes less than a clean drilled orifice of the same area. Nothing else on the page depends on it.
Compressed Air Leak Test Calculator — Decay Rate and CostBuildFigure

Why the whole system is the flow meter

A vessel of known volume at a known pressure holds a known amount of free air, and the amount is just the volume multiplied by the absolute pressure and divided by atmospheric. Drop the gauge from 120 psi to 100 psi and 120 gallons of receiver and piping has given up 20 psi worth: 16.04 cubic feet of vessel times 20 psi divided by 14.7 psia, which is 21.83 cubic feet of free air. Do it in eight minutes and the leak is running at 2.73 cfm.

Everything else on the page is bookkeeping around that one line. It is worth being clear that the atmospheric term is not decoration — dropping it, and treating the tank as holding its own volume of free air, understates the answer by a factor of about eight at shop pressures. A 120 gallon receiver at 110 psi holds 16 cubic feet of steel and 136 cubic feet of air.

The number the test comes back with

At the defaults, 2.73 cfm at 110 psi against a machine that delivers 100 cfm is 2.7 percent of the compressor gone before a single tool is picked up. At 20 kW per 100 cfm that is 0.546 kW running whenever the system is pressurised, 2,183 kWh across 4,000 hours a year and 305 dollars on the bill at 14 cents. Roughly 5.88 dollars a week.

The equivalent hole is the part that changes how people look at the shop. 2.73 cfm at 124.7 psia through openings with a flow coefficient of 0.65 is 0.00183 square inches — a single round hole 0.048 inches across, or about three sixty-fourths. Spread that over forty couplers, six hose ends, four drain valves and a dozen threaded joints and no individual leak is audible.

Running the test so the answer means something

Shut the compressor down, not just unloaded — an unloaded machine can still be venting through its own control. Leave every ball valve at every drop open, because a valve you close takes its hose, its coupler and its tool out of the measurement, and those are exactly where leaks live. Read one gauge at the start and the same gauge at the end. If a dryer purges or an automatic drain fires during the window, put the pressure it took into the correction field rather than pretending it did not happen.

Then wait. Air coming out of a compressor is hot, and a receiver cooling down after a hard afternoon loses pressure with no leak at all. Test at the start of a Monday rather than the end of a Friday and the two answers can differ by a factor of two. Running it twice, an hour apart, tells you which one you got.

Isolating where it is going

The test gives one number for the whole system, which is useful for deciding whether to care and useless for deciding where to start. Repeating it with the header valve closed splits the total between the compressor room and the shop. Repeating it again with the drops isolated one at a time splits the shop. Each run is the same arithmetic on a smaller volume, so remember to change the volume figure as well as the valve positions — that is the mistake that makes the second test look better than it is.

Questions people ask

How do I work out the volume of my compressed air system?

The receiver capacity is stamped or plated on the vessel, in gallons. Piping is the bore area times the length: a 1 inch schedule 40 pipe has a 1.049 inch bore, so 0.0060 square feet, and 200 feet of it holds 1.2 cubic feet or 9 gallons. Add any secondary or dry tank. Piping is usually a small fraction of a system with a real receiver on it, but leaving it out makes the measured leak rate come out low, so it is worth the five minutes.

Is a pressure decay test the same as the load and unload method?

No, and they answer different questions. This one shuts the machine off and watches a known volume fall, so it needs the volume and nothing else. The load and unload method leaves the compressor running with no tools in use and times the loaded and unloaded periods, which gives the leak as a fraction of delivered flow and needs a trustworthy delivered figure instead. The decay test is easier to trust on a small system because a receiver capacity is a fact printed on the vessel while a delivered flow figure usually is not.

Why does the leak rate change with system pressure?

Flow through a small sharp opening at these pressure ratios is choked, which means it is set by the absolute pressure upstream and not by how much pressure difference there is across the hole. Double the absolute pressure and you double the flow. That is why the page scales the rate measured across the test band up or down to the pressure the system actually runs at, and why dropping the header ten psi takes a real slice off the leak flow rather than nothing.

The gauge fell but I cannot hear anything. Is the test wrong?

Probably not. The default result corresponds to an opening 0.048 inches across spread over the whole shop, which is nowhere near audible over any machinery, and a single leak that size is a hiss you have to put your ear next to a fitting to find. Soapy water in a spray bottle finds them, and so does an ultrasonic detector. Before hunting, though, check the temperature question: a receiver still cooling from a hard run loses pressure on its own and reads as a leak that does not exist.

What counts as a leak rate worth fixing?

That is a money question rather than a technical one, and this page gives you the money. Multiply the rate by your specific power and your electricity price and compare it against an afternoon of somebody replacing couplers and re-taping threads. The page prints a weekly figure alongside the yearly one for exactly that reason, because a yearly number feels abstract and a weekly one does not. Nothing here says any particular rate is acceptable or unacceptable.

Related