// Sampling / Refrigerant cylinder
Taking a refrigerant sample that still means something when it arrives
9 steps, each with the reason attached: what the step protects, and what the reported number does when it is skipped. It is the procedure from our own sampling document, with the reasoning put back in.
01Arrangement
Which port, and in which order
Two valves on the source, one job each. Getting them the wrong way round is the most consequential thing that can happen at the machine, and it produces a result that looks entirely normal.
A / Source vessel
- Vapor valve
- Opens into the vapour space at the top. Used for the separate non-condensables sample, and as the dry-nitrogen inlet on low-pressure refrigerant. Not the port for this draw.
- Liquid valvein use
- Opens below the liquid level. The sample is drawn here. On a vessel with no dedicated liquid port, invert the vessel so the outlet sits under the liquid.
Short charging hose, purged before the cylinder opens
B / Sample cylinder
- Rated 400 psi
- Evacuated to 50 microns or below
- Tare weighed, tare written on the record
- Filled to 75–80% of volume, about 350 g
- Valve opened last, closed first
Stands on a scale for the whole fill — the fill is judged by weight, never by feel
Why liquid, and not vapour
6 of the 7 characteristics are liquid-phase measurements. For a blend, the liquid and the vapour above it are not the same material: the vapour is richer in whichever component boils first, so a vapour draw reports the composition of the headspace rather than the composition of the charge. Blends are judged on composition windows, so that shift is not a rounding difference — it is the verdict.
A vapour draw also leaves oil, high-boiling residue and acidity behind in the vessel, which reads on the report as a clean sample.
R-410A — how little room there is
| Component | Nominal | Allowable | Above nominal |
|---|---|---|---|
| R-32 | 50.0 | 48.5–50.5 | +0.5 |
| R-125 | 50.0 | 49.5–51.5 | +1.5 |
Percent by weight. The windows are asymmetric and two-sided, so there is no purity figure that can rescue a sample drawn from the wrong phase. Every blend window is published.
02Procedure
The draw, step by step
Follow it in order. The sequencing is load-bearing in three places: the hose is purged before the cylinder opens, the source valve closes before the cylinder valve, and the record is written at the machine rather than at the truck.
Step 1. Match the cylinder to the source, not to the plant room
Kit cylinders are rated 400 psi. Check that rating against the saturation pressure of the refrigerant you are sampling at the warmest temperature the cylinder will see — which is a vehicle or a loading dock, not the machine room you are standing in.
- Cylinder rating
- 400 psi
Why
The cylinder is the pressure vessel for the entire journey, and the pressure inside it follows the warmest hour of that journey rather than the coolest. A vessel chosen against plant-room conditions has been chosen against the least demanding part of its life.
If this goes wrong
A cylinder sized for the wrong condition is a problem in transit, where nobody is watching it and nobody can vent it.
Step 2. Evacuate the empty cylinder and prove it held
Pull the cylinder to 50 microns or below, then isolate the pump and watch the gauge. Judge evacuation on the reading, never on how long the pump ran.
- Evacuate to
- ≤ 50 microns
Why
Anything left in the cylinder is reported as though it came out of the system. Residual air is measured as non-condensables — 1.5% by volume for every blend, and per refrigerant for a single component — and residual moisture against a limit expressed in parts per million by weight. Neither can afterwards be told apart from air and water that were genuinely in the charge; there is no signature that distinguishes them. Isolating the pump also proves the cylinder and its valve hold vacuum, which is the leak check from step 7 run before you have anything to lose.
If this goes wrong
A cylinder that was pumped for a while without a gauge carries an unknown quantity of air and water into a measurement reported in parts per million. Material that left your site in specification can arrive out of it.
Step 3. Tare weigh, and leave the scale set up
Weigh the evacuated cylinder complete with its valve and cap, and write the tare on the record. The scale stays under the cylinder for the fill.
- Target fill
- ~350 g
Why
Fill is a mass measurement, and without a tare there is no way to answer either of the two questions that must be answered before you disconnect: is there enough to test, and is there too much to travel. There has to be enough material for 7 characteristics — the acidity determination alone is specified as "Titration — detection limit 0.1 ppm on a 50–60 g sample".
If this goes wrong
A cylinder filled by feel is either short, and comes back to you for a second trip, or overfilled, which is worse.
Step 4. Connect to the LIQUID valve of the source
Make the connection at the source's liquid valve. If the vessel has no dedicated liquid port, invert it so the outlet sits below the liquid level. Keep the hose as short as the job allows.
Why
6 of the 7 characteristics are defined on the liquid phase; only non-condensables is a vapour measurement. For a blend the two phases are not the same material. R-410A has a small glide, which is why its two phases get treated as interchangeable. They are not — the vapour runs richer in R-32, the more volatile component, and the R-32 window is not centred on nominal: it allows 0.5 of a point above 50.0% by weight and 1.5 below it, so the enrichment runs into the short side. A modest shift is not a modest error against 0.5 of a point. Oil, high-boiling residue and acidity barely travel in the vapour at all, so a vapour draw also under-reports those by leaving them behind in the vessel.
If this goes wrong
A vapour draw on a blend produces a plausible, internally consistent, wrong-in-a-known-direction result. That is more expensive than an obviously bad sample, because nothing about it invites a second look.
Step 5. Purge the hose before the sample cylinder is opened
With the sample cylinder still shut, sweep the line with a short burst of refrigerant at the sample-cylinder end and make the fitting up while it is still flowing. Keep the purge to the minimum that clears the hose. The sample cylinder is the last valve you open and the first one you close.
Why
The hose is full of atmospheric air, and an evacuated cylinder will draw every bit of it in the instant its valve opens. Everything step 2 achieved is undone by a hose nobody purged, and the resulting non-condensables reading is indistinguishable from air in the system.
If this goes wrong
Non-condensables that came out of your hose, judged against the 1.5% by volume limit that applies to every blend.
Step 6. Fill to 75–80% of cylinder volume, and stop
Watch the scale, not the cylinder. Around 350 g is the target for a kit cylinder. Close the source valve first, then the sample cylinder valve, so nothing continues to migrate into it after you have stopped weighing.
- Fill
- 75–80% of volume
- Target mass
- ~350 g
Why
The 20–25% you leave behind is not spare capacity, it is the compressible volume that keeps the pressure bounded when the cylinder warms up. Liquid refrigerant expands appreciably with temperature and is very nearly incompressible. While a vapour space remains, warming simply pushes liquid into it and the pressure follows the saturation curve — the curve the 400 psi rating was chosen against. Once the cylinder goes liquid-full there is no vapour left to displace, and further warming is resisted by the bulk modulus of the liquid instead. Pressure then climbs very steeply for a small rise in temperature, and it climbs with no relation to saturation pressure at all. That is hydraulic expansion, and it is why an overfilled cylinder is a hazard rather than an inconvenience.
If this goes wrong
An overfilled cylinder is a transport hazard rather than a sample, and one that cannot be opened safely has to be taken again.
Step 7. Disconnect, cap, and leak check with soap bubbles
Close both valves, disconnect, cap the outlet, then wet every joint with leak-detection solution and watch. Bubbles anywhere means the sample does not travel.
Why
A cylinder that leaks slowly does not simply arrive light. It arrives with a different composition than it left with, because what escapes from the headspace of a blend is vapour, and that vapour is enriched in the more volatile component — the same physics as step 4, running in reverse for the whole journey. A leaking vessel is also not a sealed vessel, so the moisture result becomes a measurement of everywhere the cylinder has been.
If this goes wrong
A blend that leaked reports as drifting toward its less volatile component, and there is nothing in the numbers to say that is what happened.
Step 8. Record the source temperature and the sample point
The non-condensables measurement is defined at 25.0 °C on a vapour-phase sample, so the source temperature at the time of sampling is part of the record.
Why
Non-condensables are specified as % by volume @ 25.0 °C. A measurement defined at a reference temperature is only interpretable if the condition of the source at the moment of sampling travelled with it. The sample point earns its place on the record for the same reason it does in oil work: it separates what was circulating from what had been standing.
If this goes wrong
Written up back at the truck, the source temperature becomes a recollection. Written at the machine, it is data.
Low-pressure refrigerant only
Step 9. Pressurise the source through the VAPOR valve with dry nitrogen
For low-pressure refrigerant, the source tank may be pressurised through the Vapor valve with dry nitrogen.
Why
A low-pressure refrigerant sits at or below atmospheric pressure at room temperature, so there is almost no driving pressure to move liquid into the cylinder; what transfer does start tends to flash at the fitting rather than arrive as liquid. A regulated nitrogen pad on the vapour space restores that driving pressure and moves the liquid as liquid. It goes into the vapour valve and the sample still comes out of the liquid valve, because the pad has to sit on top of the charge rather than travel with it.
If this goes wrong
Never pad a refrigerant that is subject to the non-condensables determination — you would be adding the exact thing that test measures. In practice the two rarely meet, and the reason is set out below.
03Non-condensables
This one needs a second sample
Air is the only characteristic measured on vapour, and it is measured at a stated reference temperature. Neither of those things can be satisfied by the liquid sample you have just taken.
A / Source vessel
- Vapor valvein use
- The vapour-phase sample is drawn here, with the source at rest and its temperature read and recorded at the moment of the draw.
- Liquid valve
- Closed for this draw.
Purged hose, kept short
C / Second evacuated cylinder
- Evacuated to 50 microns or below, same as the first
- Vapour fill — mass target does not apply
- Labelled as the vapour sample, not as a duplicate
Source temperature at the time of sampling goes on the record with it
Why it cannot share the liquid sample
- Phase
- vapor
- Basis
- % by volume @ 25.0 °C
- Blend limit
- 1.5%
Gas chromatography — separate vapour-phase sample. Air is not dissolved evenly through a charge — it collects in the vapour space, which is precisely the part of the vessel a liquid draw is designed to avoid. Sampling liquid and reporting non-condensables from it would be reporting a number about a region the sample never touched.
Source temperature is recorded at samplingThe non-condensables measurement is defined at 25.0 °C on a vapour-phase sample, so the source temperature at the time of sampling is part of the record.
Not every refrigerant needs this sample at all. The exemptions are in the next section, and they are not a courtesy — they follow from the boiling point.
04Low-pressure refrigerant
When the source cannot push the sample
Below atmospheric pressure at room temperature there is nothing to move the liquid with. A regulated dry-nitrogen pad on the vapour space restores the driving pressure — and the reason that is acceptable is worth knowing before you do it.
N / Dry nitrogen, regulated
- Regulator set low — enough to move liquid, no more
- Dry: the pad must not be a moisture source
Into the vapour space
A / Source vessel
- Vapor valvein use
- Nitrogen enters here, so the pad sits on top of the charge.
- Liquid valvein use
- Liquid still leaves here, from below the pad.
Liquid out, from under the pad
B / Sample cylinder
- As in the main arrangement — evacuated, tared, filled by weight
Why adding a gas does not spoil the test
It would, on the wrong fluid. Nitrogen is a non-condensable, so padding a refrigerant that is subject to that determination means adding the very thing being measured. The reason the practice is sound is that the two cases barely overlap: a refrigerant needs a nitrogen pad because its normal boiling point is at or above room temperature, and that same property is why the standard does not require a non-condensables determination on it.
Non-condensables not required — the federally incorporated text
The exemption list is not identical across editions, and the edition that applies to your work decides it. All three lists are published. If the fluid you are padding is not on the list that applies, take the vapour sample before the nitrogen goes on — or do not pad it.
05Before it ships
After the valve closes
The cylinder is full and sealed. What is left is the part that decides whether the result can be attached to anything.
Label it
Refrigerant designation, source vessel or machine, sample point, date and time, source temperature, and whether this cylinder is the liquid sample or the vapour sample. An unlabelled cylinder is a sample with no identity.
Register it
Registering before the cylinder travels means it is logged the moment it reaches the bench, and it gives you a record of what you sent and when — including anything that deviated from this procedure.
Register a sample →Ship it
The kit includes labelling. Confirm current packaging, marking and declaration requirements with your carrier before the cylinder goes out — those requirements sit with the shipper and they change.
Amalina Technologies shares ownership with Chiller Services Refrigerants, a refrigerant reclaimer in the adjacent unit, and a purity result can be an input to what reclaimed material is worth. We say so here rather than leave it to be discovered. How the laboratory is kept separate from it.