Nitrogen Pressure Test: Proving a Heat Pump Line Set Holds

What a standing nitrogen pressure test on a refrigerant line set actually demonstrates, the temperature effect that fakes a leak, and where the test pressure really comes from.

Almost everything that goes wrong with a heat pump in its first few years traces back to the refrigerant circuit being opened, joined and closed again by hand on site. The equipment arrives from the factory sealed and tested. The copper between the outdoor unit and the indoor coil does not: it is cut, bent, brazed and flared in a basement or against a wall, and every one of those operations is an opportunity to leave a path for refrigerant to escape.

A nitrogen pressure test is the step that finds those paths while they are still easy to fix. Dry nitrogen goes into the closed line set at a pressure well above anything the system will see in service, the source is shut off, and the gauge is watched. If the reading holds, the joints are sound. If it falls, something is open, and the installer would far rather know that with the wall still unfinished than three winters later when a unit is limping and the homeowner is paying for electric backup heat.

What a nitrogen pressure test proves, and what it cannot

The choice of gas is deliberate on several grounds. Nitrogen is inert, so it will not react with the copper or with the oil left in the system. It is dry, which matters because moisture inside a refrigerant circuit reacts with the oil and produces acid that quietly eats the compressor windings. And it is not a refrigerant, so a leak during the test releases nothing anyone has to account for and nothing anyone has to buy again. Compressed air fails the first two tests, because it carries water vapour with it, and oxygen fails a more serious one: oxygen and refrigerant oil under pressure are a combination nobody should create indoors.

What the test demonstrates is narrow and worth stating plainly. It shows that, at a pressure held over a period, the sealed volume did not lose gas faster than the instrument could detect. That is a statement about the joints. It says nothing about whether the line set is the right diameter, whether it was routed within the vertical separation the equipment allows, whether the insulation is continuous, or whether the system will eventually hold the correct refrigerant charge. Those are separate questions with separate checks, and a passed pressure test is regularly waved at all of them.

It also does not prove the circuit is dry. Nitrogen finds leaks; only evacuation removes moisture and non-condensables. The two steps come one after the other for that reason, and skipping the second because the first went well is a substitution that does not work.

The temperature trap that produces a false leak

The single most common misreading of a standing pressure test has nothing to do with the joints. Gas pressure in a fixed volume tracks absolute temperature, so the same sealed line set reads higher when it is warm and lower when it is cold, with no gas having gone anywhere. In practice the ratio is what matters: pressure divided by absolute temperature stays constant, so a drop in one has to be checked against a drop in the other before it means anything.

This is not a laboratory subtlety in New York. A line set pressurised in the afternoon on the sunny side of a house and read at dusk has cooled substantially, and the gauge falls accordingly. Half of it may run through an unconditioned attic or along a foundation wall where the swing is larger still. Two rules follow. Record the ambient temperature at the start of the hold and again at the end, not just the pressures. And hold the test long enough that a genuine leak separates itself from thermal drift, which for a small drop means hours rather than minutes.

The inverse error is at least as expensive: a real slow leak masked by a rising temperature. A test started in a cold morning basement and read after the heating came on can show no change at all while gas is escaping steadily. Temperature is not a correction to apply when the answer looks wrong; it is part of the reading.

How this test differs from the other checks in the sequence

Commissioning a split system involves several tests that all sound like leak checks and prove different things. Confusing them is how a system ends up signed off with a defect still in it.

That last row is the one homeowners most often hear described as “they pressurised it with nitrogen” when something quite different happened. Purging is a flow of gas through open tube during brazing, and its purpose is to keep the inside of the copper clean; scale that forms without it breaks loose later and lands in the metering device. The standing test is the opposite: everything shut, nothing flowing, a gauge and a clock. A good installation does both, at different moments, for unrelated reasons.

Where the numbers actually come from

People reasonably want a figure: how many pounds of pressure, held for how long. The honest answer is that the figure belongs to the equipment rather than to the internet. Manufacturers publish a test pressure in the installation instructions shipped with the outdoor unit, and it varies with the refrigerant the system is built around; the transition away from older refrigerants toward the current lower-global-warming blends has moved those numbers, so literature remembered from a previous generation of equipment is not a safe guide. On top of that sits whichever edition of the state mechanical code the jurisdiction has adopted, and New York City runs its own construction codes on a separate cycle from the rest of the state. Which edition governs a particular job is a question for the local building department, not an assumption to carry onto site.

Two constraints are physical rather than editorial, and they matter to anyone watching the work. Nitrogen leaves the cylinder at a pressure far beyond anything a residential line set is built to take, so a two-gauge regulator is not optional equipment, and neither is a relief device set below the weakest component’s rating. And the weakest component is often not the copper. If the indoor coil, a service valve or a filter drier is inside the tested volume, the test pressure is limited by whichever of them is rated lowest, which is why parts of the system are frequently isolated before the gas goes in.

Why this step carries extra weight on a New York retrofit

A new-construction installation runs new copper in a straight, accessible path. A retrofit rarely does. Fishing a line set through a finished house in Buffalo or Queens produces longer runs, more bends and more joints, often with sections buried in a wall or chase before anyone will see them again. Every additional brazed joint is another candidate for the test to find, and the cost of finding it afterwards rises steeply once the wall is closed – which is the practical argument for insisting the hold is complete and documented before any patching starts.

Retrofits also invite the reuse of existing copper. A line set left behind by an air conditioner may be the right size, or it may be short of the diameter the new equipment needs, and it will certainly contain the residue of an older refrigerant and its oil. Reuse is a decision about line set length and diameter limits as much as about tightness, and it is one of the few places where a pressure test that passes can still leave a system that will never perform. When a house is coming off oil heat and the whole heating load is moving onto the new equipment, that margin is not there to be given away.

The test also sets up the step that follows it. A charge is only meaningful in a circuit that holds, which is why weighing refrigerant in against the measured line set length comes after the hold and after evacuation, never alongside them. Airflow is the third leg of the same commissioning stool, and it is the one that degrades on its own over the years – a loaded blower wheel will undo a correctly charged system quietly and without any leak at all.

What to ask for as the homeowner

You do not need to supervise the work, and standing in a basement watching a gauge helps nobody. What is reasonable is to ask for the result in a form that exists after the van has gone: the pressure at the start and end of the hold, the ambient temperature at both readings, how long the hold ran, and which parts of the system were isolated during it. Those four items take a minute to write down and they are the difference between a test that happened and a test you can point at.

It is also fair to ask when the test ran relative to closing up any wall, and whether the same technician will record the evacuation result. Contractors who work regularly under the state’s programmes are used to documenting commissioning steps because the programmes ask for them, which is one practical reason to start from the state’s participating contractor listings rather than from a search result. The paperwork is not the point in itself, but a contractor who already keeps it is a contractor whose sequence is unlikely to be improvised.

Frequently Asked Questions

How long should a nitrogen pressure test be held?

Long enough that a genuine leak becomes distinguishable from temperature drift, which in a residential line set means hours rather than minutes. A very large leak announces itself immediately, but those are not the ones that cause trouble; the expensive failures are slow enough that a fifteen-minute look at a gauge cannot see them. Many installers leave the test standing overnight for exactly this reason. What makes the duration useful is recording the ambient temperature at both ends of it, because without that the pressure difference cannot be interpreted at all.

Can compressed air be used instead of nitrogen?

It should not be. Shop air carries water vapour, and moisture inside a refrigerant circuit combines with the compressor oil to form acids that attack the windings from the inside over time. Compressed air may also carry oil from the compressor that produced it. Nitrogen from a cylinder is dry and inert, which is why it is the standard for both the standing test and the brazing purge. Oxygen is a separate and far more serious mistake: under pressure with refrigerant oil present it is a hazard, not merely a poor substitute.

The pressure dropped overnight. Does that always mean a leak?

No, and this is where most false alarms come from. Gas pressure in a sealed volume follows absolute temperature, so a line set that was warm in the afternoon and cold by morning reads lower with nothing having escaped. The way to tell is to compare the pressure change against the temperature change rather than looking at the pressure alone. If the pressure fell further than the cooling accounts for, there is a leak. If it tracks the temperature, the circuit held. The same arithmetic works in reverse, and a warming system can hide a real leak entirely.

Does a passed pressure test mean the system does not need a vacuum?

It does not. The two steps answer different questions. The pressure test proves the circuit is tight; evacuation removes the moisture and non-condensable gases that entered while the tubing was open, and neither can substitute for the other. Skipping evacuation because the pressure held leaves water and air inside a system designed to contain neither, which shows up later as poor capacity, high head pressure and acid formation. The correct order is test, release the nitrogen, evacuate to a deep vacuum, verify the vacuum holds, then charge.

Is a nitrogen pressure test required, or is it just good practice?

Both the equipment manufacturer’s installation instructions and the mechanical code adopted in the jurisdiction speak to pressure testing a field-made refrigerant circuit, so it is not merely a courtesy. What they say specifically – the pressure, the medium and the documentation – depends on which edition is in force where the work is happening, and New York City administers its own construction codes separately from the rest of the state. Treat the requirement as real and the details as something to confirm locally rather than assume from a general article.

Can an existing air conditioning line set be reused for a heat pump?

Sometimes, and the pressure test is only part of what decides it. The tubing has to be the diameter the new equipment specifies, within the length and vertical rise the manufacturer allows, and it will contain residue of the previous refrigerant and its oil, which may be incompatible with what is going in. Some manufacturers require a flush or a cleaning procedure, and some conditions rule reuse out entirely. A reused set that holds pressure but is undersized will run for years while quietly costing capacity, which is the failure that is hardest to detect afterwards.