After Recovery The System Won't Evacuate: Moisture Vs Leak Decision Tree

Why this matters

You finished recovery, hooked up the vacuum pump, and the micron gauge crawls toward target and stalls, or it pulls down and then climbs the instant you valve off. After recovery the system is wet: refrigerant that just flashed out of the oil left moisture, and a system that sat open during the repair pulled in humid air. Now you cannot tell whether the slow pull is trapped water boiling off or a real leak letting outside air in. Guess wrong and you either ship a wet, acid-prone system into service or you spend an hour leak-hunting a tight system that just needed two more evacuation cycles. The decay-curve shape and a few isolation branches resolve it fast.

Symptom presentation

Two failure shapes dominate after recovery. Slow pull: the gauge takes far longer than expected to reach 500 microns, or never gets there, hovering at 1500 to 3000. Decay on valve-off: it reaches target, you isolate the pump, and the reading rises. The rise either levels off at a plateau (moisture equilibrium inside a closed system) or climbs without bound (air entering from a leak). A post-recovery system almost always shows some moisture decay on the first cycle; the question is whether that is the whole story or whether a leak rides underneath it.

Quick checks

• Pump oil first. Recovery work splashes refrigerant-laden, moisture-laden oil. Milky or low oil cannot reach deep vacuum. Change it before condemning the system.
• Core removal. Schrader cores throttle the evacuation badly. Remove cores and use core-removal tools so the system pulls through full-bore ports.
• Hose size and length. Short, large-diameter hoses pull a system down dramatically faster than the thin charging hoses left on the manifold. Restriction in the hoses mimics a slow, leak-like pull.
• Gauge at the system. Mount the micron gauge on the equipment, on a separate port from the pump, so you read the system vacuum and not the pump's.
• Ambient. Cold equipment boils off water glacially. Below ~50 F a wet system may never reach target without heat assist.

Isolation tree

After reaching the lowest vacuum the pump will deliver, valve off and watch 5 to 10 minutes.

Branch A: rises to a plateau (e.g., settles around 1000 to 2000 microns and holds). Moisture. Equilibrium between liquid water and vapor stopped the rise. Re-evacuate. Break the vacuum with dry nitrogen, hold briefly, recover the nitrogen, and pull again. This triple-evacuation sweeps water vapor out mechanically. Each cycle the plateau should drop and the rise should slow. After two or three cycles a moisture-only system holds below 500.

Branch B: rises continuously, no plateau. Air ingress: a leak. The closed system cannot reach equilibrium because the source is unlimited. Move to nitrogen pressure testing.

Branch C: cannot reach target even with fresh oil, cores out, and good hoses. Either a heavy moisture load (a flooded or rain-soaked open system) or a leak large enough that the pump and the leak reach a stalemate above target. Distinguish by isolating: valve the system off at the manifold so only the pump-and-hose side is under vacuum. If that isolated stack now drops to target fast and holds, the slow pull is the system (moisture or leak). If the isolated stack will not reach target either, your tools are the problem.

Branch D: holds well after the first re-evacuation, then degrades on a later cycle. Watch for a tool leak that opens under thermal cycling, or a Schrader port re-seated poorly. Re-check ports and core caps.

Confirming diagnosis

• Standing vacuum. A system holding below 500 microns for 15 minutes with negligible rise is dry and tight. This is the acceptance criterion.
• Nitrogen pressure decay. If Branch B or C points to a leak, recover any residual, pressurize with dry nitrogen to the nameplate/AHRI test pressure, and time the pressure. A pressure drop over an hour confirms a real leak; a stable pressure clears the system and points the slow pull back to moisture.
• Curve documentation. Asymptotic curve equals moisture; linear unbounded curve equals leak. Log both axes.

Remediation

• Moisture: triple-evacuate with nitrogen sweeps; replace the liquid-line filter-drier (and suction drier if fitted) on any system that was open to atmosphere or shows persistent wet behavior. A saturated drier and trapped water form acids that destroy compressors.
• Leak: locate by nitrogen pressure test with electronic detector and bubbles, re-braze under flowing nitrogen to avoid internal oxidation, re-test, then re-evacuate to a clean standing hold.
• Restriction artifact: new pump oil, cores removed, large-bore short hoses, gauge at the system. Re-run the decay test once the tooling is corrected; many "won't evacuate" calls are pure tooling.
• Always: finish with a documented standing vacuum below 500 microns before charging by weight.

References

• EPA Section 608 Technician Certification (recovery completeness, refrigerant handling, no venting)
• AHRI Standard 700, Specifications for Refrigerants (moisture and contaminant limits)
• ASHRAE / ACCA evacuation and dehydration guidance (triple-evacuation method, micron acceptance, standing-vacuum hold)
• Equipment manufacturer installation literature for nameplate evacuation target and nitrogen test pressure