How to Run a Standing Nitrogen Pressure Test on a New Line Set

Why this matters

A vacuum pull tells you the system holds against about one atmosphere pushing in. A refrigerant system in service holds hundreds of psi pushing out. A braze joint with a pinhole, a flare that was not quite torqued, or a Schrader core seated crooked can pass a decay test and leak the first hot afternoon, because 15 psi of push from the outside does not open what 350 psi from the inside does. The standing nitrogen test is the check that loads every joint in the direction it will be loaded for the next twenty years, before any refrigerant is at risk.

It is also the cheapest leak to find. Before the charge goes in, a leak costs a soap bottle and a re-braze. After, it costs a recovery, a re-braze, a new evacuation, and a callback the customer counts.

Safety first: this is stored energy

A full nitrogen cylinder holds over 2,000 psig. Never connect a cylinder to a system without a pressure regulator, and never use a regulator without a relief device on the outlet side; one opened valve without a regulator puts cylinder pressure on a line set rated for a fraction of it. Chain or strap the cylinder upright, keep the valve cap on whenever it moves, and stand to the side of a valve when opening it.

A joint that fails at test pressure throws fragments and oil at whoever is close. Safety glasses go on before the first stage, and nobody puts their face to a joint while pressure is rising. Soap and look from the side.

Nitrogen displaces oxygen without any smell or warning. Vent test pressure outdoors, never into a closet, crawlspace or attic where someone is working, and do not vent a large volume into a small room. Stored pressure in the line set is released fully, to 0 psig on the gauge, before any core is pulled or any vacuum hose goes on.

The rig

 N2 CYLINDER        REGULATOR           MANIFOLD
 over 2,000 psig -> outlet set at or -> high side hose
 chained upright    below test press.   gauge reads here
                    relief on outlet         |
                                             v
 OUTDOOR UNIT                    port on LIQUID valve
  liquid service valve: CLOSED   (line set side)
  suction service valve: CLOSED  port capped, or a
        |                        second gauge on it
        |                              |
   liquid line                   suction line
        |                              |
        +-------- INDOOR COIL ---------+
        every braze joint and flare soaped

The outdoor unit's service valves stay closed, so the test loads only what the crew built: the line set, the indoor coil and every joint between them. The factory charge in the condenser stays where it is.

Start state

All brazing done under a nitrogen purge per Brazing With a Nitrogen Purge Standard, joints cooled to the touch, flares torqued, the TXV or EEV in its installed position. A digital gauge with 1 psi resolution, a pipe clamp thermometer, leak detection solution that does not freeze or dry fast, and the test pressure written down before anything is connected.

Find the test pressure. It is the lowest of these: the test pressure on the outdoor unit's nameplate or in its installation instructions, the maximum on the indoor coil's label or manual, and any lower limit stated for an accessory in the circuit. The shop never exceeds it. In the worked example below, both documents state 450 psig, so 450 psig is the ceiling and also the target.

Steps

1. Set the regulator low and pressurize to the first stage, about 150 psig. Open the cylinder, then raise the regulator slowly while watching the manifold gauge, and shut the manifold valve at 150 psig. A gross leak, a missed joint or a cap left off a port shows up here at a pressure that cannot hurt anyone. Going straight to full pressure finds the same leak with more energy behind it.

2. Soap every joint. Every braze joint, every flare, both service valve ports, the Schrader cores, and any union the crew did not make but is now inside the test. Watch each one for a full minute; a slow leak grows a bubble, it does not foam on contact. If one bubbles, stop: vent the pressure outdoors to 0 psig, repair, and start over at step 1.

3. Raise to the second stage, about 300 psig, and soap again. Some leaks open only under load. A flare that sealed at 150 can weep at 300 as the nut and the face move.

4. Raise to the test pressure, and no higher. Set the regulator to the test pressure itself, so a valve left open cannot carry the system past it. Shut the manifold valve, close the cylinder, and disconnect or isolate the cylinder from the manifold so a slow regulator creep cannot add pressure during the hold.

5. Record the starting reading. Pressure on the gauge, line set temperature from a clamp on the copper, and the time. The temperature is of the pipe, not the air. On a line set that crosses an attic and an outdoor wall, clamp it where most of the length runs and note where; that is the temperature the gas mostly sits at.

6. Hold for the shop's period. A common starting point is at least 30 minutes after the reading settles on a short, fully visible run, and overnight on a long run or one with concealed joints. Settling matters: the first minutes after a fill show the gas cooling from compression, and a drop then is heat leaving, not gas.

7. Read again and correct for temperature. Pressure in a sealed volume moves with absolute temperature, so a line set that cooled overnight reads lower without losing a molecule. Convert to absolute, correct, convert back:

• absolute pressure = gauge pressure plus 14.7 psi at sea level (use local barometric pressure at altitude)
• absolute temperature in Rankine = F plus 460
• expected end pressure (absolute) = start absolute times end Rankine divided by start Rankine

At these pressures that works out to a little under 1 psi per degree F. Compare the reading with the expected value, not with the starting value.

8. Release and hand off to evacuation. Vent the nitrogen slowly, outdoors, to 0 psig. Only then pull cores and set up the vacuum rig; see related: Vacuum Rig Layout for a Fast Clean Evacuation. The service valves stay closed until the evacuation and decay test pass per Evacuation and Vacuum Decay Verification.

Worked example: an attic run that leaked at stage one and passed overnight

A 3-ton heat pump: air handler in an attic, 46 ft of line set, one coupling mid-attic on the suction line where two coils were joined, two braze joints at the air handler stubs and two at the outdoor service valves. Test pressure from the manual and the coil label: 450 psig.

Stage one, 150 psig. Soap at the mid-attic coupling grew a bubble at the base of the fillet within about 20 seconds. The crew vented to 0 psig outdoors, cleaned and re-brazed the joint under a nitrogen purge, let it cool, and started over. Second time at 150 psig: every joint dry for a full minute.

Stage two, 300 psig. All joints dry.

Test pressure. 450 psig at 2:30 pm, regulator set at 450 and the cylinder closed and isolated. Clamp on the suction line in the attic: 92 F.

Next morning, 7:30 am. Gauge 430 psig. Clamp at the same spot: 68 F. That is 20 psi below the start, and a crew reading it raw would be hunting a leak.

The correction:

 start   450 psig + 14.7 = 464.7 psia
         92 F + 460      = 552 R
 end     68 F + 460      = 528 R
 expect  464.7 x 528 / 552 = 444.5 psia
         444.5 - 14.7      = 429.8 psig
 read    430 psig
 diff    430 - 429.8       = 0.2 psi, inside 1 psi
                             gauge resolution: PASS

The whole 20 psi drop is the attic cooling 24 F overnight. Checked from the other direction: at the starting condition the gas moves 464.7 divided by 552, or about 0.84 psi per degree, and 0.84 times 24 F is about 20.2 psi, the same drop the full correction gives.

What a leak would have looked like. Had the gauge read 412 psig at 68 F, the reading would sit about 18 psi below the 429.8 psig expected, with no temperature left to explain it. That is a leak, and it goes back to step 2 at test pressure with soap, starting at the joints the crew made and then the ones it did not.

The finish test

The test passes when the corrected end pressure matches the expected pressure within the gauge's resolution after the full hold, and every joint stayed dry at each stage. Write the four readings (start pressure and temperature, end pressure and temperature), the times, the expected value and the result onto the job record, then vent and evacuate. A stable reading before any vacuum is pulled is the gate; nothing gets evacuated on a test that has not closed.

What changes the call

A run in direct sun. A line set on a south wall can swing 30 F or more between a cloud and full sun inside a short hold. Either clamp where the sun reaches and read when conditions are steady, or hold overnight and read at dawn, when the whole run is near one temperature.

A mini-split with flare connections. Flares are the joints most likely to weep at test pressure and least likely to show it at a low stage. Soap them at every stage, and check that the indoor unit's manual test pressure is not lower than the outdoor unit's. See related: How to Make and Torque a Flare on a Mini-Split Line Set.

An analog gauge. A gauge marked in 5 psi divisions cannot resolve the difference between a pass and a slow leak on a short hold. Either use a digital gauge or extend the hold until the expected leak would move the needle a full division.

A reused line set. The same test applies, with the line set capped off and tested alone first; see related: How to Check an Existing Line Set Before Reusing It.

References

• Outdoor unit nameplate and installation instructions, and indoor coil label: field test pressure
• See related: Vacuum Rig Layout for a Fast Clean Evacuation
• See related: Brazing With a Nitrogen Purge Standard
• See related: Evacuation and Vacuum Decay Verification
• See related: How to Check an Existing Line Set Before Reusing It