Deep Vacuum and Micron Gauge Evacuation
A compound gauge that reads "30 inches" tells you almost nothing. Here is how to pull a real deep vacuum, hit 500 microns, and prove the system is dry and tight with a decay test before you ever release the charge.
Charging after the vacuum holds?
Dial in superheat and subcooling to verify the charge once refrigerant is in the system.
In This Guide
Why a Deep Vacuum Matters
Evacuation does two jobs at once: it removes non-condensable gases (mostly air and the nitrogen you can't sweep out) and it boils off moisture that got into the lines during brazing, storage, or an open system. Both are quiet killers.
Air trapped in a system doesn't condense in the condenser like refrigerant does. It collects at the top of the coil, raises head pressure, drives up discharge temperature, and steals capacity. Moisture is worse. Combined with the heat of compression it attacks POE oil, forms acids, corrodes copper and bearings, and freezes at the metering device — you'll chase a "low charge" that is really a plug of ice cycling open and shut at the TXV or orifice.
POE oil is the reason microns matter now
The mineral oil in old R-22 systems tolerated a little moisture. The POE (polyolester) oil used in R-410A, R-454B, R-32 and other HFC/A2L systems is hygroscopic — it grabs moisture out of the air aggressively and holds onto it. That is exactly why deep vacuum discipline went from "nice to have" on R-22 to non-negotiable on modern equipment.
Microns, Not Inches of Mercury
Your manifold's compound gauge tops out around 30 inches of mercury (inHg) of vacuum and simply pins there. But that last needle-width covers an enormous range. A micron gauge measures absolute pressure electronically, in microns of mercury, where atmospheric pressure at sea level is 759,968 microns and a perfect vacuum is 0.
| Microns | Roughly inHg | What it means |
|---|---|---|
| 760,000 | 0" | Atmospheric — system open to air |
| 25,000 | ~29" | Where a compound gauge already looks "pinned" |
| 5,000 | ~29.7" | Free water still boiling — not dry yet |
| 1,000 | ~29.9" | Getting close; decay ceiling for many techs |
| 500 | ~29.9" | Industry target — dry and tight |
| 250 | ~29.9" | Excellent; common spec on critical work |
Notice the whole useful range — 5,000 down to 250 microns — lives inside the last two-tenths of an inch on a compound gauge. That is the entire argument for a micron gauge. You cannot manage what you cannot measure, and the compound gauge simply cannot measure it.
The number to memorize
Pull to 500 microns, isolate the pump, and confirm the system holds below roughly 1000 microns on a decay test. That is the widely accepted AHRI-aligned benchmark for a clean, dry system.
The Tools That Actually Get You There
A slow, disappointing evacuation is almost always a tooling problem, not a pump problem. The refrigerant gauge manifold you charge with is the single worst thing you can pull a vacuum through — the hoses are narrow, the depressor pins choke flow, and the rubber outgasses. Fix the path and everything speeds up.
- A two-stage vacuum pump — a CFM rating matched to system size (5–6 CFM covers most residential and light commercial). More important than raw CFM is clean oil.
- Fresh vacuum pump oil — old oil is saturated with moisture and will cap your ultimate vacuum. Change it before the job, not after.
- Core removal tools — these let you pull the Schrader valve cores while under pressure so you evacuate through full-bore openings, then reinstall cores without losing the vacuum.
- Short, large-bore vacuum-rated hoses — 3/8" or larger, kept short. Standard 1/4" charging hoses can double or triple your pull-down time.
- A micron gauge tied into the system itself — never read microns at the pump, where the pump makes its own suction look better than the system really is.
- Dry nitrogen — for the pressure test beforehand and for a nitrogen sweep if the system was badly contaminated.
Evacuate from both sides
Pulling a vacuum only through the suction port makes the metering device a bottleneck — the liquid side can stay wet for hours. Connect to both the liquid and suction service ports so you evacuate the whole system evenly. On systems without a liquid-side port, a nitrogen sweep before final evacuation helps push moisture out of the dead legs.
The Evacuation Procedure, Step by Step
Step 1 — Pressure test first, always.
Charge the system with dry nitrogen to the appropriate test pressure and confirm it holds. A vacuum is a terrible leak test — atmosphere pushes in through a leak, so a leaky system can still look like it's pulling down.
Step 2 — Recover any nitrogen, then remove the cores.
Bleed off the test nitrogen and install core removal tools on the liquid and suction ports. Back the cores out so you have full-bore flow.
Step 3 — Connect short 3/8" vacuum hoses to both sides.
Tie the micron gauge into the system through its own valved port, isolated from the pump so you can valve off for the decay test.
Step 4 — Fresh oil, open the gas ballast, start the pump.
The ballast helps carry water vapor out of the pump early in the pull-down; close it near the end for the deepest vacuum.
Step 5 — Pull down toward 500 microns.
Watch the micron gauge fall. If it stalls around 5,000 microns and sits there, that plateau is water boiling — give it time or apply gentle heat to the accumulator/receiver; do not use a torch.
Step 6 — Confirm 500 microns with the pump still running.
Reaching 500 with the pump open only proves the pump can outrun what's in the system. The real test comes next.
Step 7 — Valve off and run the decay test.
Isolate the system from the pump and read the rise (covered below).
A2L caution (R-32, R-454B)
A2L refrigerants are mildly flammable. Vacuum pumps are not rated for flammable atmospheres, so make sure the system is fully recovered and the space is ventilated before evacuating. Follow the equipment manufacturer's A2L service procedure and keep ignition sources away from the service ports.
Ventora — Your AI HVAC Companion
PT charts, superheat and subcooling calculators, EPA 608 code answers, and AI troubleshooting. Free on iOS.
The Decay Test: Reading the Rise
Hitting 500 microns is only half the story. Isolate the system from the pump, let the micron gauge sit, and watch what the pressure does over the next several minutes. The shape of the rise tells you exactly what you're dealing with.
Holds flat
Rises a little, then stabilizes and stays put — say 500 climbing to 700 and holding. System is dry and tight. You're cleared to charge.
Rises then levels
Climbs to around 1,500–3,000 and then plateaus. That's remaining moistureboiling off. Go back on the pump and keep dehydrating.
Climbs without stopping
Rises steadily and never levels off, heading back toward atmosphere. That's a leak. Stop, fix it, and re-test — don't charge it.
Worked example: was it moisture or a leak?
A 3-ton R-410A split pulls down to 480 microns. You valve off the pump and log the gauge:
t = 0:00 → 480 microns (pump isolated)
t = 1:00 → 900 microns
t = 3:00 → 1,650 microns
t = 8:00 → 1,780 microns
t = 15:00 → 1,800 microns (holding)
The rise slowed and leveled near 1,800 rather than marching to atmosphere. That's the signature of moisture, not a leak. The fix is more evacuation time — and often a fresh liquid-line filter drier — not chasing a phantom leak with a bubble solution.
If instead that same system had gone 480 → 1,200 → 4,500 → 12,000 → 40,000 and kept climbing, you have a real leak somewhere the nitrogen test missed (or a loose core depressor / gauge connection). Break it down and find it before another minute of pump time.
Why It Won't Pull Down
When a system stubbornly refuses to reach 500 microns, work the list from the pump outward before you assume the system is the problem:
| Symptom | Likely cause | Fix |
|---|---|---|
| Pump alone won't hit deep vacuum | Saturated pump oil | Change the oil; blank-off test the pump |
| Painfully slow pull-down | 1/4" hoses & cores still in | Core tools + short 3/8" vacuum hoses |
| Stalls near 5,000 microns | Free water boiling off | More time, gentle heat, nitrogen sweep |
| Liquid side stays wet | Evacuating from one side only | Connect both service ports |
| Reads great at pump, bad in decay | Micron gauge sensor at the pump | Move the gauge onto the system |
A quick sanity check: with the pump running and the micron gauge on the pump's inlet, a good pump and good oil should reach the pump's ultimate vacuum (often well under 50 microns) within a minute or two. If the pump can't do that on its own, no amount of system fussing will help.
Releasing the Charge
Only after the decay test holds do you commit refrigerant to the system. The sequence protects your hard work:
- Reinstall the valve cores through the core removal tools without breaking the vacuum.
- Break the vacuum by admitting refrigerant vapor — never let air back in.
- Weigh in the nameplate charge with an accurate scale, adjusting for line-set length per the manufacturer's data.
- Verify the final charge with superheat (fixed orifice) or subcooling (TXV) once the system stabilizes.
Field rule
A system that hit 500 microns and held a flat decay is dry and tight. If it wouldn't hold, no charge and no gauge reading will save it — the moisture or leak you skipped will come back as an acid burnout or a callback. The vacuum is where reliability is won.
Ventora — the AI HVAC Assistant in Your Pocket
From PT charts and micron targets to superheat, subcooling, and EPA 608 code answers — get instant, field-ready help and AI troubleshooting right from your phone.