The Cylinder That Was Not What the Label Said

Why this case is worth reading

A recharge that makes the gauges look worse instead of better sends a tech hunting through the system for a fault, and every fault he can name is in the system. The cylinder on the scale is the one component nobody suspects, because it came sealed and labeled. This case walks the hypotheses in the order a good tech would test them, and the test that finally settled it is one every HVAC tech already knows, pointed at something they rarely point it at.

The safety frame for the whole call: refrigerant work here was done by an EPA Section 608 certified technician, with gloves and eye protection against liquid refrigerant burns, gauges and hoses rated for the system's pressures, and, once the gas was in doubt, no ignition sources near the unit or the recovery setup.

The readings that did not add up

A 3-ton split system, R-410A per its nameplate, 88 F outdoor air on a late-summer afternoon. On arrival it was undercharged by every sign: suction 112 psig, superheat 24 F, head 330 psig, subcooling 3 F against a nameplate target of 10 F. The tech found the leak at a Schrader core on the suction service valve, replaced the core with a core removal tool without opening the system, and confirmed it tight with leak solution.

He then added 2.0 lb as liquid, metered slowly into the suction side, from a new 25 lb cylinder the shop had bought two weeks earlier from an online seller offering it at a noticeably low price. Fifteen minutes later, reading saturation from a standard R-410A chart:

• Suction 136 psig, saturation about 47 F; suction line 56 F, so superheat 9 F.
• Head 470 psig, saturation about 130 F; liquid line 112 F, so subcooling 130 minus 112, 18 F.
• Condensing over ambient 130 minus 88, 42 F. This unit's charging data expects roughly 15 to 25 F.

High head, high subcooling, low-normal superheat. On the face of it, an overcharge, two pounds after a reading of undercharged.

First hypothesis: he overshot the charge

This is the honest first guess, and it fits the readings. Overcharge stacks liquid in the condenser, which raises subcooling and raises head pressure, and with a slow leak nobody knows how much refrigerant was actually missing, so 2.0 lb could be too much.

What argued against it was proportion. Twenty minutes earlier subcooling was 3 F, and 2.0 lb took it to 18 F, 15 F of rise, while the condensing split went to 42 F, nearly double the top of the expected band. An overcharge large enough to do that would be more than a modest overshoot from undercharged. Suspicious, not conclusive. He held off recovering any charge, because recovering refrigerant to fix a reading you do not yet understand trades one unknown for another.

Second hypothesis: a restriction

A restriction in the liquid line or at the metering device starves the evaporator, so it drives suction pressure down and superheat up, while liquid backs up behind the restriction and subcooling rises. The rising subcooling fit. The rest went the wrong way: suction had risen from 112 to 136 psig and superheat had fallen from 24 to 9 F. A restriction cannot produce higher suction and lower superheat. Ruled out by direction, not by inspection.

Third hypothesis: the condenser cannot reject heat

A dirty coil, a failed fan motor or a recirculating discharge raises head pressure on any charge. The tech checked it directly: coil clean on both faces, fan turning the right way at its nameplate amps, the unit 3 ft from the nearest wall with nothing over it. Ruled out by observation.

Fourth hypothesis: noncondensables, and where they could have come from

Noncondensables are gases (usually air, sometimes nitrogen) that do not condense at system pressures. They collect in the condenser and add their own pressure to the refrigerant's, so head pressure reads high, the saturation temperature read off the chart from that pressure is inflated, and subcooling computed from it reads high too. That fits all three readings.

The problem was the source. Noncondensables get in when a system is opened and poorly evacuated, and this one was never opened: the leak repair was a core swap under pressure. The only thing that had entered the system in the past hour was the 2.0 lb from the cylinder. That moved the question.

The test that moved the question to the cylinder

The standing pressure check: with the system off and its refrigerant settled, the pressure inside a container of a single refrigerant (or a near-azeotropic blend like R-410A) equals the saturation pressure at the temperature of the coldest point holding liquid. If the reading is higher than the chart at that temperature, there is something in there that is not that refrigerant.

He shut the system down and let it sit 30 minutes. The coldest point in the system was the indoor coil in a 72 F house. He set the new cylinder and a second R-410A cylinder from the shop's usual distributor side by side in the shade for an hour, both at 78 F by a probe taped to each shell, and read both on the same gauge:

STANDING PRESSURE CHECK   chart = R-410A saturation
item             temp F   gauge psig  chart psig  over
system, at rest  72 coil  236         208         +28
cyl A, new       78       268         228         +40
cyl B, shop      78       229         228         +1

Each "over" is gauge minus chart: 236 minus 208 is 28, 268 minus 228 is 40, 229 minus 228 is 1.

Cylinder B, from the known channel, read within 1 psi of the chart on the same gauge, which rules out the gauge and the thermometer. Cylinder A read 40 psi over at the same temperature. The system read 28 psi over: the contaminant was diluted into a larger existing charge, so its excess was smaller than the cylinder's, which is the direction it should run. And it settles the first hypothesis: an overcharged system at rest still reads the chart at its coldest point, because more refrigerant is still only refrigerant. Only something that is not R-410A puts the standing pressure above it.

A refrigerant identifier, the instrument that samples a gas and reports its composition, confirmed it the next morning: cylinder A's sample read below purity for R-410A, with a large share flagged as air and a smaller share it could not identify. That last part mattered more than the air.

What was done, in order, and why that order

The cylinder came out of service first. It was tagged, bagged and moved to the shop's quarantine bin with photos of its label, valve, lot and shell before anything else, so a rushed recharge could not reach for it again. See related: The Suspect Counterfeit Quarantine SOP.

The whole system charge was recovered, not vented and not partly bled. Venting refrigerant during service is prohibited under the Clean Air Act Section 608 rules at 40 CFR 82.154, and a contaminant does not change that; it just changes where the recovered gas goes. The unidentified share is the reason the recovery was handled as possibly flammable. Counterfeit refrigerant has been found to contain methyl chloride (R-40), which is flammable and reacts with aluminum to form compounds that ignite in air, and an identifier that cannot name a component cannot rule it out. So: recovery done outdoors, no brazing torch lit, no smoking, the recovery machine one rated for flammable refrigerants, and the gas sent into a dedicated recovery cylinder labeled "contaminated, unknown component", never into the shop's shared recovery cylinder, where it would have spoiled everything else in it. Had the shop owned no machine rated for flammables, the right move was to stop with the system sealed and call the reclaimer for instructions.

The system was restored to a clean baseline. New liquid-line filter drier, a nitrogen pressure test, evacuation to 500 microns with a 10-minute decay check holding under 1,000 microns, then the nameplate charge weighed in by scale from cylinder B. On restart the tech stood clear of the service valves and hoses as the compressor came up, then read the result: head 365 psig (saturation about 110 F, a 22 F split over the 88 F air, inside the unit's band), suction 128 psig, superheat 11 F, subcooling 10 F on target. Every reading now agreed with every other one, which is the real sign the charge was right: the earlier readings had each been plausible alone and impossible together.

The supplier and the report

The recovered cylinder went to a reclaimer with the identifier printout, and the reclaimer's receipt went in the job file. The shop wrote to the online seller with the lot number, the standing-pressure table and the identifier result, and asked where the lot came from; the seller refunded without answering. The owner reported the seller to EPA, which takes tips on suspected illegal hydrofluorocarbon imports and sales, and noted a second flag: the seller had never asked for a Section 608 certification, which the sales restriction in 40 CFR 82.154 expects before selling refrigerant for this kind of equipment.

Refrigerant now comes only from the shop's approved distributors; see related: The Safety-Critical Parts Purchasing SOP. And every new cylinder gets the same two-minute check before its first use: a standing pressure against the chart at a measured shell temperature, beside a known-good cylinder when one is on the truck.

References

• EPA, 40 CFR Part 82 Subpart F (Section 608 refrigerant management: venting prohibition and sales restriction)
• EPA guidance on recognizing and reporting illegal HFC imports and sales
• Manufacturer charging data and R-410A pressure-temperature chart for the equipment serviced
• See related: Counterfeit and Contaminated Refrigerant
• See related: The Suspect Counterfeit Quarantine SOP