1,500 Microns and a Vapor Charge: Two Mistakes on One Freezer
A 120 lb R-404A walk-in freezer rack had its evaporator replaced. The technician triple-evacuated with dry nitrogen sweeps and pulled the system to 1,500 microns, isolated the pump for five minutes, saw the reading hold, and broke vacuum. He then charged vapor from an upright R-404A cylinder until the sight glass cleared, and set the final charge by watching suction pressure. Ambient is 85°F, the system has a TXV and a receiver, and the nameplate charge is 120 lb. The store opens in the morning and he wants to leave.
1. Evaluate the evacuation. Was 1,500 microns with a five-minute hold an acceptable standard, and what should have been done instead?
Read the worked answer
Marginal at best, and the hold test was too short and too loose to prove anything.
Reasoning:
- The working target for a system with a receiver and a fresh evaporator is 500 microns, measured at the system with a micron gauge - not at the pump, and not on a manifold gauge.
- The decay test is what actually proves the job: isolate the pump and watch the rise for 10–15 minutes. A slow rise toward 1,000–1,500 microns suggests moisture still boiling off (or a small leak); a rise that keeps climbing toward atmosphere is a leak, and no amount of pumping will fix it.
- At 1,500 microns the system is roughly three times the recommended moisture threshold; any residual water will freeze at the TXV on a freezer and show up as a callback in a week.
What should have happened: if the decay test failed, break vacuum with dry nitrogen and repeat the sweep - that is the whole point of triple evacuation. Nitrogen breaks the vacuum, dilutes the water vapor, and carries it out on the next pull. Log the final micron level and the decay result on the work order; it is the only evidence the evacuation was real.
2. The technician charged R-404A as a vapor from an upright cylinder. Why is that wrong, and how should the charge have been introduced?
Read the worked answer
R-404A is a zeotropic blend (R-125 / R-143a / R-134a) with a small but real temperature glide. Its components have different boiling points, so vapor drawn off the top of the cylinder is richer in the more volatile components than the liquid below.
Consequences of vapor charging:
- The system receives the wrong composition - lower capacity, different pressures, and a charge that cannot be corrected by adding more of the same vapor.
- The remaining liquid in the cylinder is now off-spec too, so the next job gets a bad charge.
Correct method: charge liquid from the cylinder. Take liquid from a dip-tube (liquid) valve directly into the liquid line while the system runs. If liquid must enter on the suction side, meter it through a throttling valve or a charging manifold so it flashes before reaching the compressor - never slug liquid into a compressor.
Note: this is a blend-specific rule. A single-component refrigerant such as R-22 can be vapor-charged; a zeotrope cannot.
3. The sight glass cleared and the suction pressure "looked right." Is that sufficient to sign off the charge? How should the charge actually be set?
Read the worked answer
No. On a blend system, a clear sight glass and a plausible suction pressure are weak evidence and can both be satisfied by a wrong charge.
Do this instead:
- Weigh it in. The nameplate charge is the starting point; add for the actual line length and any receiver/piping changes from the evaporator job. Record the scale weight.
- Verify with subcooling at the condenser outlet (TXV system): subcooling tells you whether the receiver has enough liquid behind the valve. Rising subcooling with high head pressure points to overcharge or non-condensables.
- Cross-check superheat at the evaporator to confirm the TXV is feeding correctly and the evaporator is fully active.
- Confirm it is actually R-404A with a refrigerant identifier if there is any chance the previous company used something else - mixing is the one error you cannot charge your way out of.
Judgment: the sight glass tells you about liquid seal, not about charge weight or composition. On a 120 lb rack, weighing is fast, repeatable, and defensible; pressure-watching is a guess that the callback will disprove.