HVAC Vacuum Pump & Evacuation Guide

By Jennifer Kirstens · August 30, 2026 · 8 min read

PES Supply is expanding into refrigerant service tools — including vacuum pumps and micron gauges. If you'd like early access, contact us to join the waitlist. This guide walks pump sizing, evacuation procedure, the 500-micron standard, moisture testing, and A2L-ready equipment considerations.

Every gram of non-condensable (air) or moisture (H₂O) left in a refrigerant system compounds into service problems: elevated head pressure, acidic breakdown of POE oil, TXV strainer icing, and premature compressor failure. On A2L systems (R-32, R-454B) moisture is even worse — the higher operating pressures push more water into solution, and acid formation attacks brushless-motor windings that were never designed for it. The single biggest quality lever a residential HVAC tech has is a proper deep evacuation to 500 microns with a documented decay test.

500 microns of mercury (roughly 0.066 kPa absolute) is 1/1520 of atmospheric pressure. At that vacuum level water boils at about 30 °F — which is why a proper evacuation actively drives moisture out of the system as vapor. Anything above 1000 microns and the water simply sits there, mixed into the oil.

A vacuum pump's "3 CFM" or "6 CFM" rating is its free-air displacement at atmospheric pressure. As the pump pulls down, actual mass removal drops rapidly — a 6 CFM pump is moving well under 1 CFM by the time you're at 500 microns. But the free-air rating still predicts how long the evacuation will take, because the bulk of the time is spent moving the initial air load out.

Table 1 — Vacuum pump CFM to system tonnage

Table 2 — CFM to line-set volume rule of thumb

The system's internal volume (line-set + coils + accumulator) drives evacuation time more than tonnage does. Use this as a sanity check.

A single-stage pump has one rotary-vane chamber; a two-stage pump has two in series. The second stage extends the ultimate vacuum from around 75 microns (single) to 15 microns (two-stage). For a 500-micron target that seems like overkill — but the second stage also gives you a much steeper pull-down curve, so a 6 CFM two-stage will hit 500 microns 30–40% faster than a 6 CFM single-stage on the same system.

Buy two-stage. Single-stage pumps are legacy tools for pool and refrigeration-truck work.

The AHRI Guideline B, ACCA Manual RS, and every major OEM install manual converge on 500 microns as the minimum acceptable evacuation for a residential split-system A/C or heat pump. The reasoning: at 500 μm and 70 °F ambient, virtually all liquid water has boiled off, and the residual non-condensable partial pressure is trivial relative to a full charge.

Where you go deeper than 500:

• 300 microns: Mitsubishi hyper-heat, LG Multi V VRF, and some Fujitsu Halcyon models call for 300 μm with a 30-minute stability window.
• 250 microns: Semi-hermetic supermarket rack systems using POE oil.
• 50 microns: Ultra-low-temp cascade freezer service. Rare in HVAC but a real category.

Table 3 — Micron target by application

Skipping steps here is the #1 cause of comeback service calls on new installs. The industry-standard procedure:

Reading a decay test is where junior techs make mistakes. Three failure signatures every field tech should recognize:

Signature 1 — Moisture

The system pulls down to 500 μm, you isolate, and the reading climbs steadily to 1500 μm then levels off. This is water vapor. Water boils under vacuum at low temp, and residual water will continue to boil off until vapor pressure equalizes. Continue evacuating; the curve will flatten each subsequent cycle.

Signature 2 — Leak

You isolate at 500 μm and the reading climbs continuously — 800, 1200, 2000, 5000, atmospheric. No leveling. This is a leak, either in the system or in your service connections. Recheck fittings first, then leak-test the system.

Signature 3 — Trapped refrigerant

You isolate at 500 and the reading jumps to 3000+ in the first minute, then hovers there. This is trapped liquid refrigerant behind a closed service valve (very common on VRF and mini-split installs where the tech forgot to open the service valve on the outdoor unit). Open the valve, re-evacuate.

For A2L systems with POE oil, or any system with a known refrigerant leak that pulled atmosphere in, single-evacuation may not remove enough moisture. The triple-evacuation procedure:

1. Evacuate to 1000 μm.
2. Break vacuum with dry nitrogen to 5 psi. Hold 5 min.
3. Evacuate to 500 μm.
4. Break with nitrogen to 5 psi again. Hold 5 min.
5. Evacuate to 500 μm and hold decay test.

Each nitrogen break dilutes remaining moisture 10× or more. Three cycles reduces water content by roughly 1000× vs a single-pass evacuation.

Vacuum pumps don't hold refrigerant under pressure, but for A2L work look for:

• Brushless motor — no brush-arc ignition risk near residual refrigerant vapor exhausted from the pump.
• Sealed on/off switch and gas-ballast valve.
• Isolation valves at the intake port — required for holding vacuum during hose swaps.
• Bluetooth micron gauge that you can place at the far service port and read from outside the equipment room.

Table 4 — Pump feature checklist by tier

Vacuum pump oil is a light distillate that absorbs moisture and refrigerant. As it saturates, the pump's ultimate vacuum degrades. Rules for pro use:

• Change oil every job on wet systems (visible moisture, active leak).
• Change oil every 3–5 jobs on clean-tie-ins.
• Change oil at the start of every day for VRF and commercial work.
• Warm the pump up 3 minutes before pulling vacuum — cold oil is thicker and pulls slower.
• Use the gas ballast valve for the first 10 minutes of any evacuation — vents water vapor before it condenses in the oil.

The pump can only pull vacuum as fast as the hose diameter allows. A standard 1/4" SAE charging hose has roughly 1/4 the conductance of a 3/8" evacuation hose, and 1/9 the conductance of 1/2". For serious evacuation work — anything over 3 tons or over 25 feet of line — use 3/8" hoses at minimum, and 1/2" for VRF and commercial.

Even better: remove the Schrader cores during evacuation using a core-removal tool (Appion MegaFlow, JB Industries CRT-165). This eliminates the primary conductance restriction and can cut evacuation time by 40%.

Every reputable pump manufacturer publishes a spec sheet with more than just CFM. The important secondary numbers:

• Ultimate vacuum (microns): The lowest pressure the pump can theoretically reach with a closed intake. 15 μm is excellent, 25 μm is good, 75 μm is single-stage territory.
• Micron pull-down time to 1000 μm on a 1-liter reservoir: A more honest speed metric than CFM. Look for 60 seconds or less on a mid-tier pump.
• Motor horsepower: 1/2 HP is standard for 6 CFM two-stage. 3/4 HP for 8 CFM. Motor size drives duty cycle — undersized motors overheat and drop performance on long jobs.
• Oil capacity: 300–600 mL for pro-grade pumps. Larger reservoirs tolerate more moisture load before oil-change is required.
• Duty cycle: Continuous is standard for pro pumps. Cheaper units have 30 minute-on / 30 minute-off ratings that make VRF work impossible.
• Noise level: Under 70 dB(A) matters for indoor and rooftop work. Above 80 dB(A) puts you in required-hearing-protection territory.

Do not buy on CFM alone. A 6 CFM pump with 25 μm ultimate and a saturated-oil pull-down time of 3 minutes will lose every evacuation race to a 4 CFM pump with 15 μm ultimate and a 45-second pull-down time.

Micron gauges use one of two sensor technologies: thermistor (older, cheaper, drifts with oil contamination) or Pirani (newer, more stable, more expensive). Thermistor cells need to be cleaned or replaced every 40–80 hours of use; Pirani cells last thousands of hours. For A2L work and warranty documentation, Pirani sensors are the pro-tier choice.

Calibration is done against a reference — either a sealed reference cell (built into the Fieldpiece SVG3, Testo 552i) or against atmospheric pressure at power-on. The atmospheric self-calibration works if the gauge is at sea level; at elevation you need a barometric-compensated model or you'll read consistently off by 50–150 microns.

Even with correct equipment, evacuations fail. Below are the field-diagnosed causes techs run into most.

The pump won't pull below 1500 microns

This is almost always saturated oil. Change the oil, warm the pump for 3 minutes, and try again. If the pump still won't pull below 1500 with fresh oil on a known-good system (short 1/4 inch line, direct to the gauge), the pump is worn — rotor-to-housing clearance is out of spec and vanes need replacement. Second most common cause is a stuck-open gas-ballast valve.

The reading drops fast then plateaus at 800–1000 microns

Almost always a slow leak at a manifold fitting or Schrader core — not the system itself. Swap in a known-good hose, verify Schrader depressor is fully sealed, and rerun. If the system still won't pull below 1000 with clean fittings, do a bubble test on the outdoor service valves.

Pump pulls to 400 microns but decay jumps to 5000+ in 1 minute

Trapped liquid refrigerant behind a closed service valve. Very common on VRF/mini-split installs where the outdoor-unit service valve was left closed after brazing. Open the service valve fully and re-evacuate. Also check for a closed txv, closed liquid-line service valve, or plugged filter-drier.

Pump pulls fine but reading swings wildly (500 to 900 back to 500)

Micron gauge sensor contamination. Oil got into the thermistor cell during a prior job. Pull the sensor, wipe with isopropyl (99%), let dry 5 minutes, and re-test.

Evacuation takes 3× longer than expected

Most likely 1/4 inch charging hoses instead of dedicated 3/8 inch or 1/2 inch evacuation hoses. Second most likely: Schrader cores not removed. Third: excessively long or coiled hose runs. Straighten hoses and keep them as short as reasonable.

Every proper evacuation starts with a nitrogen pressure test and, on marginal jobs, uses nitrogen breaks between evacuation cycles. Dry nitrogen (99.998% purity, dew point under -80 °F) is inert, non-condensable, and displaces refrigerant safely during brazing. Rules for nitrogen work:

• Always use a pressure regulator. Bottle pressure is 2200–2400 psi; you're testing systems rated for 550 psi max. A blown fitting from unregulated nitrogen can kill you.
• Test at 150 psi low side / 550 psi high side for R-410A, R-454B, and R-32. R-22 and R-134a test lower (150/250).
• Hold for 30 minutes minimum. Longer holds catch slower leaks.
• Correct for ambient temperature swing — 4 °F change is roughly 1 psi. A digital manifold with temperature compensation is worth it here.
• Do not use house-air or oxygen. Oxygen with refrigerant oil is an explosion risk; house air is loaded with water vapor.

A 20 lb / 80 cu ft nitrogen bottle from a welding supply typically runs $30–$50 refill and lasts a solo residential contractor two to three months.

For OEM warranty and service-manager review, capture the decay-test result. Modern wireless micron gauges (Accutools BluVac+ Pro, Fieldpiece SVG3, Testo 552i) log to phone apps and produce PDF reports directly.

• Final micron reading before isolation.
• Decay rise over 10 minutes (start, +5 min, +10 min).
• Ambient temperature (moisture boil-off is temperature-dependent).
• Pump model and CFM used.
• Hose diameter used.
• Whether cores were removed.
• Any nitrogen-break cycles (triple-evac).

This documentation trail is what OEMs ask for on warranty compressor failure claims. Skip it and Daikin, Carrier, Trane, and Lennox will deny — they assume improper evacuation until proven otherwise.

Water vapor pressure is temperature-dependent. At 70 °F ambient, water boils at roughly 500 microns. At 40 °F, it boils at closer to 3200 microns — meaning your 500-micron target on a cold-weather install is not actually driving water off effectively. Two strategies for cold-weather evacuation:

• Warm the system. Apply gentle heat (heat gun on low, or wrap the accumulator with a heating pad) to raise the internal temperature to 60 °F+ before final decay.
• Triple-evacuate. The nitrogen-break dilution reduces total moisture load and compensates for reduced boil-off.

In hot ambient (95 °F+) the reverse can bite you: outdoor pump housing runs hot, oil viscosity drops, and the ultimate vacuum degrades. Keep the pump shaded and check oil at mid-day.

The evacuation procedure above applies identically to R-410A, R-454B, R-32, R-134a, and R-22 systems. PES Supply stocks these refrigerant SKUs:

• R-410A Refrigerant — 25 lb Cylinder
• R-134a Refrigerant — 30 lb Cylinder
• R-22 Refrigerant — 30 lb Cylinder (Reclaimed, AHRI 700)

For a deeper walk-through of the whole service toolkit see the HVAC Refrigerant Service Tools Buying Guide. For leak-testing procedure that precedes every evacuation, see Refrigerant Leak Detection Methods & Tools.

A complete pro-grade evacuation kit — 6 CFM two-stage pump, 3/8 inch evacuation hoses, wireless micron gauge with Pirani sensor, core-removal tools, and vacuum pump oil — runs $850 to $1,400 depending on brand tier and whether you include an A2L-listed pump variant. PES Supply is expanding into refrigerant service tools. Vacuum pumps, micron gauges, evacuation hoses, and core-removal tools — A2L-ready options included. Contact us to join the waitlist.

What is the industry-standard vacuum level for HVAC systems?

500 microns of mercury, with a decay test showing ≤500 μm rise in 10 minutes after pump isolation. This is the ACCA Manual RS and AHRI Guideline B standard for residential and light commercial split systems.

Do I need a two-stage vacuum pump?

Yes. Single-stage pumps top out around 75 microns and pull down 30–40% slower. Two-stage pumps reach 15 microns ultimate and are the pro-tier standard.

What size vacuum pump for a 3-ton residential system?

4–6 CFM two-stage is the right size. 6 CFM is faster and future-proofs you for slightly larger jobs; 4 CFM is fine if your work is mostly residential.

Where do I place the micron gauge during evacuation?

At the far end of the system from the vacuum pump — typically on the opposite service port. Reading at the pump gives an artificially low value that doesn't reflect the actual system vacuum.

How long should an evacuation take?

Residential 3-ton system: 25–50 minutes to reach 500 microns with a 6 CFM pump. Add 15 minutes of decay testing after isolation. Total 40–65 minutes.

Can I use 1/4" hoses for evacuation?

You can, but you shouldn't. 3/8" or 1/2" evacuation hoses have 4–9× the conductance and cut evacuation time roughly in half. On any system over 3 tons, 3/8" is a hard minimum.

What is triple evacuation and when do I need it?

Evacuate to 1000 μm, break with dry nitrogen, evacuate to 500 μm, break with nitrogen again, then evacuate to 500 μm and hold. Required when a system had a known atmospheric contamination event or on POE-oil systems that have been open to atmosphere for more than a few hours.

How often should I change vacuum pump oil?

Every job on wet or contaminated systems. Every 3–5 jobs on clean tie-ins. Daily on VRF and commercial. Cold, saturated oil is the #1 cause of not being able to hit 500 microns.