What it is

On a ductless install the factory charge lives in the outdoor unit behind closed service valves. Your job is to connect a clean, dry, leak-free line set and indoor head, prove it holds pressure, get all the air and moisture out, and only then release that charge into the system. Skip or rush any step and you trap moisture, non-condensables, or a leak inside a sealed system — and now you own a problem that EEVs and inverter boards hate.

Two separate tests, in this order: a nitrogen pressure test to find leaks, then a vacuum to dehydrate and de-air. They prove different things. Vacuum alone does not find leaks reliably; pressure alone does not dehydrate.

How it works

Dry nitrogen is inert, cheap, and holds steady pressure, so a drop on the gauge over time means a leak — the gas isn't condensing or being absorbed, it's escaping. That makes nitrogen the right tool to prove tightness.

A deep vacuum boils off any water inside the system (water boils at room temperature once you get the pressure low enough) and pulls out air. Air is a non-condensable: trapped in a running system it raises head pressure and kills capacity. Moisture combines with POE oil and refrigerant to form acids that eat the compressor. So you evacuate to dehydrate and de-air — not to leak test.

The micron gauge is how you actually see vacuum quality. A regular compound gauge can't resolve the difference between "good enough" and "system full of moisture still boiling."

In the field

Order of operations on a typical single-zone install:

  1. Finish and torque all flares, indoor and outdoor, with the service valves still closed.
  2. Pressure test with dry nitrogen. Connect through the service-port side, charge the line set and indoor coil up to test pressure. Don't pressurize the outdoor unit's sealed side — its valves stay shut, protecting the factory charge.
  3. Hold and watch. Let it sit and compare the gauge against the starting pressure, correcting for temperature swing (pressure rises in sun, falls at night — roughly a couple psi per 10°F). A true leak keeps dropping regardless of temperature.
  4. Find any leak with bubbles or an electronic sniffer at every flare, repair, and re-test. Don't move on until it holds.
  5. Release the nitrogen, then connect the micron gauge at the port (not on the manifold body — put the micron gauge as close to the system as you can, ideally on the suction service port, with the core removed for speed).
  6. Pull a vacuum with a good pump and large-bore (core-removed or low-loss) connections. Small hoses and Schrader cores choke evacuation and double your time.
  7. Reach target, then isolate and decay-test. Valve off the pump and watch the micron gauge.
  8. Open the service valves to release the factory charge into the system only after the vacuum passes.
  9. Add trim charge for line-set length over the factory-included length, per the data plate.

Normal values & targets

  • Nitrogen test pressure: for R-410A/A2L line sets, techs commonly test in the 300–500 psig range (some specs call out a specific value like ~550 psig; never exceed the lowest-rated component, and never exceed the equipment's stated limit). Hold for a meaningful interval — 15 minutes to several hours; longer catches slow weeps.
  • Vacuum target: pull to 500 microns or lower. Many techs shoot for 300–500 microns before the decay test.
  • The 500-micron decay test: isolate the pump and the system should not rise above roughly 500 microns in the first minute or two, and should stabilize. A slow creep that levels off around 1000–1500 microns usually means residual moisture still boiling. A fast, unbounded climb means a leak — go back to the pressure test.
  • Trim charge: typically specified as grams of refrigerant per foot (or meter) of line beyond the included length — read the data plate. Common figures land around 12–20 g per foot of line over the base, but use the actual number for that unit.

Common faults & what they mean

  • Vacuum won't get below ~1500 microns and won't decay-pass: moisture in the system or a tiny leak. If it climbs and never levels off, it's a leak. If it climbs slowly and plateaus, it's moisture — keep pulling, maybe sweep with a small nitrogen charge and re-evacuate.
  • Pressure test drops overnight: real leak. Temperature can't account for a steady multi-psi loss. Re-soap every joint.
  • System runs high head after startup: air left in from skipping or shortcutting evacuation (non-condensables), or overcharge.
  • Acid/dark oil on a young system: moisture was left in. The evacuation was inadequate.

Tech tips & gotchas

  • Never pressure-test or purge with oxygen or compressed shop air. Oxygen plus oil is an explosion. Shop air carries moisture, defeating the whole point. Dry nitrogen only.
  • Put the micron gauge on the system, not on the pump, and isolate the pump for the decay test — otherwise you're reading the pump's vacuum, not the system's.
  • Remove the Schrader cores (with a core-removal tool) during evacuation. Cores cut your effective flow path to a pinhole and can triple evacuation time.
  • A cold day fights your vacuum. Moisture boils off slower when everything's cold; warm the indoor coil area or be patient.
  • Don't trust a needle gauge for vacuum. It'll read "29 inches" while the system is still full of moisture. The micron gauge is the only honest reading.
  • Triple evacuation (pull down, break with nitrogen, repeat) is the move on a system you suspect got wet or sat open.

Safety / code notes

  • Always regulate nitrogen through a pressure regulator with a relief; a raw cylinder is 2000+ psig and will burst a line set instantly.
  • Eye protection during pressure testing and evacuation.
  • A2L equipment: follow the listing for test medium, charge limits, and leak-check method; the principles here apply but honor the equipment-specific instructions.