Deep Evacuation and Weigh-In Refrigerant Charging

Why this matters

Moisture and non-condensables in a refrigerant circuit cause acid formation, copper plating, expansion-valve hunting, and compressor motor burnout. The only repeatable way to remove them is a deep vacuum verified with a micron gauge isolated from the pump. Eyeballing the manifold's analog "vacuum" gauge is not evacuation; that gauge cannot resolve below 29 in/Hg, which is still 25,000 microns, which is still wet. A correctly-evacuated and weigh-in-charged system protects the compressor warranty and gives the unit the rated capacity stamped on the data plate. A pinched-off "topped-off-by-pressure" charge does neither.

Equipment list

• Two-stage rotary vane pump, 6 CFM minimum for residential, 10 CFM for light commercial. Pump oil clean and at the full mark. Change pump oil after every job that pulled significant moisture; saturated oil cannot pull deep vacuum.
• Digital micron gauge with a fresh battery. Acceptable instruments include Fieldpiece SMAN/SVG3, Yellow Jacket 69086, AccuTools BluVac. Analog vacuum gauges are not acceptable.
• Vacuum-rated hoses with core depressors removed, or large-bore (3/8 in or 1/2 in) evacuation hoses with quarter-turn ball valves. Standard 1/4 in refrigeration hoses with Schrader depressors choke vacuum flow by an order of magnitude.
• Heated digital refrigerant scale, 220 lb capacity, 0.05 oz resolution. Cheap scales drift; verify with a known weight before each job.
• Manifold with isolation valves at the gauge port and pump port.
• Replacement Schrader cores and core-removal tool (Appion MGAVCT or equivalent). Always replace cores after evacuation; the old cores have been off-gassed for hours and reseat poorly.

Evacuation procedure

1. Recover any existing refrigerant per EPA 40 CFR Part 82 before evacuating; never vent. Recovery is a separate step with its own machine.
2. Pull both Schrader cores using a core remover with isolation valves in place. This is the single highest-leverage improvement most techs skip. Cores restrict evacuation conductance by 80 to 90 percent.
3. Connect the evacuation rig: pump - manifold - large-bore hoses - core tools - service ports. Connect the micron gauge as far from the pump as possible (at the suction service port or on a tee at the condenser), never at the pump intake. Reading at the pump shows the pump's vacuum, not the system's.
4. Open all manifold and isolation valves. Start the pump. Watch the micron gauge fall.
5. Pull to 250 microns or lower. The reason 500 is the industry line is that water's boiling point tracks pressure: at atmospheric it boils at 212 F, at roughly 4,600 microns it boils at 32 F, and by 500 microns it is boiling at well below zero F. Anything under 500 means free water in that circuit is being actively driven off rather than sitting there. Pulling to 250 gives you margin to rise into during the decay test.
6. Close the isolation valve between system and pump (decay test). Watch the micron gauge for 5 minutes.
   • Stable below 500 microns: system is dry and tight; proceed to charge.
   • Rises and stabilizes between 1,000 and 3,000 microns: moisture remaining. Continue pumping.
   • Rises continuously past 5,000 microns and keeps climbing: active leak. Re-pressurize with nitrogen and find it before charging.

The pass criterion is the one in step 6: isolated, the reading rises and then stabilizes, and it stabilizes below 500 microns. A rise from 250 to 400 that then holds flat is a pass; that is trapped gas equalizing, not moisture and not a leak. A reading that keeps climbing with no plateau is a leak no matter how low it started. Anything else, fix before charging. The micron gauge is the verdict, not the pump runtime.

Weigh-in charging

Blends must be charged as liquid, and the reason is fractionation. A blend sitting in a cylinder has a vapor space richer in the lighter component, so drawing vapor off the top pulls that component out preferentially and leaves both the cylinder and the system off-composition. R-410A, R-454B, R-454C, and R-407C are all blends and all get charged liquid, every time, with the cylinder upright on the dip tube.

R-32 and R-22 are single-component refrigerants and cannot fractionate. R-32 is difluoromethane, R-22 is chlorodifluoromethane; there is nothing in either to separate. Charge them liquid anyway, because it is faster and because weigh-in is the method regardless, but do not repeat the fractionation reason for them. Knowing which of these is a blend and which is not is also what tells you whether a partial charge loss can be topped off or whether the circuit has to be recovered and weighed in fresh.

1. Confirm the OEM charge spec. Read the unit data plate. For a split system, the plate charge applies up to a published line-set length (typically 15 to 25 ft); add 0.6 oz per foot of liquid line above that, per manufacturer (Carrier, Lennox, Trane all publish charge-adjustment tables for their line sets).
2. Reinstall the Schrader cores. Pull a brief vacuum on the manifold and hoses (do not break the system vacuum yet).
3. Place the upright cylinder on the heated scale. Tare. Connect liquid port (the line marked LIQUID, typically dip-tube equipped on modern cylinders).
4. Open the cylinder slowly, charge liquid into the high side with the system OFF. Watch the scale. When the equalized charge stops flowing, close the cylinder valve.
5. If the full charge will not equalize into the high side (common on heat-pump units with bidirectional valves), start the unit in cool and continue weighing in through the suction service port using the Carrier-style "throttle the liquid through the manifold to flash it" technique. Never feed liquid straight into a running suction; throttle through the manifold valve so it flashes before reaching the compressor.

Verification

After full nameplate charge is in, run the unit at design conditions for 15 minutes. Measure superheat at the suction service valve (TXV systems: target 8 to 12 F; piston systems: cross-reference the manufacturer's charging chart). Measure subcooling at the liquid line (target 8 to 12 F at design conditions, OEM-specified). Both should land in the OEM window with the weighed charge. If they do not, the unit has a real problem (restriction, contamination, mismatched components); do not adjust the charge to band-aid bad superheat / subcool readings.

References

• EPA 40 CFR Part 82 Subpart F: Refrigerant Recycling
• ACCA Standard 9-2010: HVAC Quality Installation Specification
• AHRI Guideline B: Reference Guide for Recovery, Recycling, and Reclamation
• Carrier Service Manual 25HCC Charging Procedures, Section 5
• Lennox Service Application Note SAN-09 (Deep Evacuation)
• ASHRAE Handbook, HVAC Systems and Equipment 2024, Chapter 1 (Refrigerant Handling)