Refrigerant Line Sets, Evacuation, and Charging
Everything between the outdoor unit's service valves and the indoor coil is your work, and it is unforgiving. A line set that is one size too small, a braze made without nitrogen, or a system charged on top of a shallow vacuum will run for a season and then fail in a way that looks like a compressor problem. This chapter, the third of four on installation, covers sizing the piping, brazing it cleanly, proving it does not leak, evacuating it properly, and putting the correct amount of refrigerant in it. The commissioning arithmetic of superheat and subcooling is developed further in the refrigeration chapters; here we focus on the installation sequence and the mistakes that get built into a system on day one.
Sizing the Line Set
Liquid and suction line diameters come from the manufacturer's line-sizing tables for that specific equipment, indexed by capacity, total equivalent line length, and vertical separation. They do not come from whatever diameter is left on the truck, and they do not come from matching the stub-outs on the old unit. Undersizing the suction line is the single most common line-set error, and its signature is a system that is fine on a short run and disappointing on a long one: excessive pressure drop in the vapor line lowers suction pressure at the compressor, which lowers density, which lowers mass flow and therefore capacity, while raising compressor discharge temperature. Vertical geometry adds two separate problems that candidates routinely confuse. When the condenser sits above the evaporator, the liquid line has to lift liquid refrigerant, and that vertical lift costs pressure, which eats into subcooling and can cause the liquid line to flash before the metering device. When the evaporator sits above the condenser, the tall suction riser becomes an oil-return problem: refrigerant oil circulates with the refrigerant and has to be carried back up to the compressor by vapor velocity, so the riser must be sized to maintain velocity and a properly formed trap is placed at the base of the riser to collect oil and let the vapor push it up in slugs.
Brazing With Nitrogen, and Why Scale Matters Later
When copper is heated in the presence of air, the inside surface oxidizes and forms a black flaky copper oxide scale. Flowing dry nitrogen through the tubing during brazing displaces the oxygen and prevents that scale from forming. The correct flow is a low, steady trickle, just enough to displace air; a strong flow chills the joint and blows the filler metal out of the fitting. The reason this matters is not cosmetic and it does not show up on the day of installation. Once the system runs, refrigerant flow eventually breaks the scale loose, and it travels to the two smallest passages in the circuit: the filter drier and the metering device. A plugged screen on a TXV or a fixed orifice produces low suction pressure, high superheat, and a starving evaporator, and it is diagnosed as a refrigerant problem by technicians who did not know the system was contaminated at birth. The same nitrogen cylinder is used for pressure testing, always through a regulator fitted with a relief device, because full cylinder pressure is many times the working pressure of the equipment and a failure with the regulator bypassed is a serious injury. Brazing in an attic adds the fire hazard: clear combustibles, use a fire barrier behind the joint, and maintain a fire watch after you finish.
The Commissioning Sequence: Test, Evacuate, Then Charge
The order is fixed and the exam tests it directly. After the last braze, you pressure test the line set and coil with dry nitrogen to find leaks. Only after the system holds pressure do you evacuate, because pulling a vacuum on a leaking system tells you nothing and wastes an afternoon. Evacuation removes two things a vacuum pump can remove and a purge cannot: non-condensable gases such as air and nitrogen, and moisture. Depth is measured with a micron gauge, not with the compound gauge on a manifold, which cannot resolve anything meaningful below about 29 inches of mercury. The widely used field target for a residential split system is 500 microns or lower, held after the pump is valved off. That decay test is where the diagnostic value lives. If the gauge falls to 400 microns, and after isolating the pump it rises and then levels off around 2,500 microns, the system is not leaking to atmosphere; moisture inside is still boiling and adding vapor until it reaches equilibrium. If instead the pressure rises continuously without levelling, you have a leak. Evacuate through both service ports with large-diameter hoses and, where possible, with the valve cores removed, because a quarter-inch hose and a Schrader core are severe restrictions at deep vacuum and will make a good pump look weak.
Putting the Charge In
Modern split systems ship with a factory charge in the condenser sized for a stated line-set length, and the manufacturer specifies additional refrigerant per foot beyond that base allowance because the extra tubing adds internal volume. Mini-split systems, especially multi-head systems, make this explicit in their instructions. After evacuation, the service valves are still front-seated, holding the factory charge in the outdoor unit; opening them fully releases that charge into the line set and coil. The defensible way to arrive at the final charge on a new TXV system is to weigh in the calculated charge with a scale, then verify by measuring liquid line subcooling against the manufacturer's target. On a fixed-orifice or piston system there is no subcooling target to charge to, because the metering device does not hold a constant superheat; those systems are charged to a target superheat read off the manufacturer's charging chart, which takes indoor wet-bulb and outdoor dry-bulb as inputs. Two refrigerant-handling details matter here. A zeotropic blend, which is most modern refrigerants, must be removed from the cylinder as liquid and metered through the gauge set so it flashes before reaching the compressor; charging vapor from the cylinder removes the more volatile component first and changes the composition of both the cylinder and the system. And a crankcase heater on a unit installed in a cold climate should be energized for several hours before the compressor is first started, so liquid refrigerant that migrated into the oil boils out before startup.
Filter Driers, Flare Joints, TXV Bulbs, and Support
A liquid line filter drier removes moisture and traps particulate before it can reach the metering device, and it is directional. Installed backwards, debris that the outlet screen has collected can wash off in the reverse direction and travel straight into the metering device. Mark the flow arrow and check it. A drier with a measurable temperature drop or frost across it in operation is not doing its job; it is restricting flow, and that produces low suction pressure with high superheat and a cold spot just downstream of the drier. Mini-split service valves use flare connections rather than brazed joints, and a reliable flare is made by cutting square, deburring the tube, forming a clean flare with the correct tool, and torquing the nut to the manufacturer's specification. Overtorquing splits the flare; undertorquing leaks slowly for a year. On a system with a thermostatic expansion valve, the sensing bulb location controls whether the valve can do its job. On suction lines larger than seven-eighths of an inch, the bulb is mounted on the side of a horizontal run, at roughly the four or eight o'clock position, so it senses vapor temperature rather than the liquid or oil that runs along the bottom of the tube, and it must be clamped tightly and insulated. Finally, support horizontal refrigerant and drain lines at regular intervals with hangers sized so they do not crush the suction line insulation.
Last updated: September 2026