A new heat pump or air conditioner can be rated at any efficiency you like, but if it slowly leaks refrigerant it will lose capacity, run longer, and eventually fail. According to ACHR News, installers and trainers see many startup leaks tied to two skipped steps: not flowing nitrogen during brazing, and not proving the system tight before evacuation. Both steps use the same inexpensive gas and a handful of tools.
Who this helps: homeowners comparing install quotes, technicians standardizing their process, and builders writing commissioning requirements.
Job one: flowing nitrogen while brazing
When copper is heated to brazing temperature with air inside, the inner wall oxidizes and forms a dark scale. HVAC School explains that this scale weakens joints and can break loose, and ACHR News notes that it mixes with refrigerant oil and can clog metering devices, filters, and screens. The fix is to displace the air with dry nitrogen. HVAC School describes a brief purge at roughly 20 to 100 standard cubic feet per hour (SCFH), followed by a low "flow" of about 2 to 5 SCFH during the braze, with an open end so pressure never builds. NYC HPD makes flowing nitrogen through the tubing during brazing a written requirement for its projects.
The tool that matters here is a flow meter (or a regulator with a flow scale), not just a pressure regulator. Pressure alone cannot tell you how much gas is moving through an open line.
Job two: the standing pressure test
After the joints are made, the system is filled with dry nitrogen, isolated from the cylinder, and watched. A pressure drop means a leak. The PNNL Building America Solution Center heat pump commissioning guide describes residential test pressures typically between 250 and 600 psig, held for about 30 to 60 minutes, and stresses that the manufacturer's recommended pressure comes first. NYC HPD sets a target of 550 psig or higher if the manufacturer specifies it, with a one-hour minimum for line sets under 200 feet and 24 hours for longer runs such as most VRF and multi-zone systems. For commercial refrigeration, EPA leak-tight installation guidance calls for at least 300 psig and a 24-hour stand with no more than a 1 psi change on the same gauge. With A2L refrigerants, ACHR News reports that a standing test of at least one hour is normally required.
Temperature can fake a leak, or hide one. Nitrogen pressure rises and falls with temperature. NYC HPD tells installers to measure outdoor shade temperature at the start and end of the test and correct by about 1 psi per °F. A line set that cools 10°F in the evening can lose roughly 10 psi without any leak at all.
Tool checklist
| Tool | What it does | What to look for |
|---|---|---|
| Dry nitrogen cylinder | Inert gas for purge, flow, and test | Dry, oil-free nitrogen. Never oxygen or shop air. |
| High-pressure regulator with relief | Controls test pressure | Range that covers the manufacturer's test pressure, plus a relief valve |
| Flow meter or flow regulator | Sets purge and braze flow | Readable in SCFH at low flow (2 to 5 SCFH) |
| Digital pressure gauge or probe | Watches for decay | NYC HPD specifies 0.1 psi resolution and does not allow analog gauges |
| Thermometer | Temperature correction | Shade reading at start and end of the test |
| Bubble solution or leak reactant | Pinpoints a leak once pressure drops | Applied to every braze, flare, valve, and core |
Where press-connect fittings fit in
Mechanical press fittings skip the torch entirely, which removes the oxide problem. ACCA notes that the Uniform Mechanical Code (2018 and 2021 editions) allows press-connect fittings for copper refrigerant joints, provided they comply with UL 207. They still need the same standing pressure test. A press joint is a different way to make a connection, not a reason to skip proving it.
A homeowner's quick check
- Ask whether nitrogen will flow during brazing, or whether press fittings will be used.
- Ask for the test pressure, hold time, and start and end readings in writing.
- Ask whether the test pressure came from the manufacturer's installation instructions.
- Ask that the evacuation (micron) reading be recorded after the pressure test.
Limitations
A pressure test proves tightness on the day of the test. It cannot predict vibration damage, rub-outs, or corrosion years later. Test pressures and hold times vary by manufacturer, system size, and local requirements. The numbers above are typical published guidance, not a substitute for the installation manual. Hold times under an hour can miss very slow leaks, which is one reason high-resolution digital gauges have become common.
Not an endorsement: this article names no tool or fitting manufacturer. Choose equipment by range, resolution, pressure rating, listing, and serviceability. HVAC Zone Inc is a multi-brand dealer, and the Hub is brand-neutral.
Planning a new system or chasing a repeat leak? Request a consultation with HVAC Zone for a site-specific assessment.