Aircon compressor oil: the lubricant that has to travel
A compressor needs lubricant, and part of that lubricant cannot stay where it is needed. It leaves with the gas, goes right around the loop, and comes back only if the flow is quick enough to carry it. The pipe route decides that.
By Team Snowflake | Updated 8 Aug 2026
What the oil does, and why none of it can be sealed in
Compressor oil holds three jobs at once, and all three happen inside the shell. It keeps a film between bearing surfaces that would otherwise touch. It seals the small gaps the machine compresses against. It draws heat away from the parts that generate it. Oil sitting anywhere else in the system is doing none of those things.
Keeping it in is not an option, because the gas being pumped passes through the same space the oil occupies. A fine mist leaves on every discharge, too slight to notice and too steady to ignore. Allied Air Enterprises' refrigerant piping design guidelines state that small amounts leave the compressor constantly and circulate with the refrigerant throughout the system. What leaves has to be brought back, or the bearings and contact surfaces lose the film they run on.
That turns the pipework into two circuits sharing one set of pipes. One moves heat, which is the job the system was bought to do. The other carries lubricant out to whichever room sits furthest away and brings it home, which nobody raises at the quotation stage. The second one is what the design rules in an installation manual are quietly protecting.
None of this counts as a fault, and none of it is unusual. A healthy machine does it continuously for as long as it runs. What differs from one installation to the next is whether the return trip finishes.
The class of oil arrives with the compressor
Which generic class sits in the sump follows from the refrigerant the machine was built around. HVAC School's reference on refrigerant oils sets out the families in use. Mineral oil and alkylbenzene go with the older CFC and HCFC refrigerants. Polyol ester goes with the HFC and HFO refrigerants in current equipment, and polyvinyl ether is an alternative that is common in ductless and VRF systems. None of that is an owner's decision or an installer's preference.
The class matters here for one reason. That same reference names three things the return trip depends on: how readily the oil and refrigerant mix, how thick the oil is, and how fast the refrigerant travels. The first two were settled by whoever built the compressor. The third is settled by whoever ran the pipe, on site, on the day. Polyol ester also takes on water readily, and where that leads belongs with moisture in the refrigerant circuit rather than here.
How the oil gets back: swept along, not drained
Oil does not find its own way to the compressor. It gets dragged there, pushed along the inside of the pipe by refrigerant vapour moving over it. The trade word is oil entrainment, and it describes a film clinging to the tube wall that the gas has to peel forward.
Speed is the entire mechanism, which is why published piping guidance is written as a velocity rather than as a pipe size. Allied Air Enterprises' guidelines set a minimum of roughly 800 feet per minute through horizontal suction lines. Vertical suction risers are given 1,200 feet per minute as the minimum and 1,500 as the preference, whatever the height of the riser happens to be.
Gravity is the whole of the gap between those two figures. On a flat stretch the gas only fights the film's grip on the copper. On a climb it fights the grip and the weight together, so the same oil wants half again as much speed to keep moving. A vertical rise is therefore treated as its own design problem, not as extra length. How far that rise is allowed to go is set by the height difference between units published for the model.
Two habits in the same document follow from this, and neither is decoration. A trap belongs at the foot of a vertical rise. Oil gathers there until the gas has enough behind it to shift the load up in one movement, and taller lifts get further traps spaced along the riser. The run should also carry no sags or low points. Anything that dips collects oil and keeps it, which is one reason hard copper is preferred to soft coil on long horizontal stretches.
Return outranks efficiency wherever the two disagree. Those guidelines note that the suction line's most important function is bringing oil home, and that on very long runs a higher pressure drop may have to be accepted to hold the velocity up. A designer will knowingly give away some capacity rather than let the film stop moving. That trade never appears on a quotation, and it is the clearest statement of how the priorities are ordered.
Why an oversized pipe is a fault, not generosity
A larger suction line carries less oil, not more. The same mass of gas through a wider bore travels slower, and the film that depends on that travel stops being carried. Allied Air Enterprises' guidelines put it directly: an oversized suction line can drop refrigerant velocity below what is needed to bring oil back to the compressor.
The pipe wall is where this bites, and the wall is the slowest part of any pipe. That document notes that the larger the pipe, the higher the velocity has to be at the centre to hold a given velocity at the wall surface. Oil is not riding in the middle of the gas stream. It is stuck to the surface, exactly where the flow is weakest.
Oversizing on a real job is seldom a decision. It is the size that was on the van, or a round-up because bigger sounded safer, or copper reused from the system that came out. The reuse case is the common one in a flat being refitted, and it is the version that looks most like thrift. Pipe sized for the old machine is wrong for the new one whenever the new one moves less gas than the old one did.
Nothing about a wide line gives itself away on handover day. It holds pressure, it passes a vacuum test, it cools the room. The material is right, the joints are sound, and the only thing wrong with it is a dimension nobody will ever measure again. Total run and vertical drop carry published ceilings for closely related reasons, and the maximum pipe length for a given model is where those are written down.
Liquid lines can be oversized too, with a different consequence. A wider liquid line holds more refrigerant, and the same guidelines connect a larger system charge to slugging and to oil dilution. Dilution is a question about a machine that has been sitting, not one that is running, and that argument is made elsewhere.
What breaks the return on pipework that was sized correctly
A correctly sized run brings oil back at the flow it was sized for, and not below it. Anything that lowers the flow lowers the carrying force with it, and the pipe has no way to make up the difference.
Part load is the first and the largest of those conditions. Allied Air Enterprises' guidelines work through an example where a line chosen for full capacity falls to roughly 850 feet per minute once the system drops to its minimum output. That is enough along horizontal runs and not enough up the vertical risers. Flat stretches keep working while the climb stops, which is the order those earlier figures predict.
The double riser exists to answer this, and it says more about the priority than any rule of thumb could. Two risers run side by side with a trap between them. At low output the trap fills until it seals the second riser off, the gas is forced up one pipe instead of two, and the speed returns. An installation will absorb extra fittings, extra joints and extra cost to keep a film of lubricant moving. Nothing else in the pipework earns that.
Inverter equipment sits at part load by design, which changes what the sizing has to satisfy. A machine bought because it modulates will run under its rated output for most of the time it is on, so the bottom of that range is the operating point the pipe has to serve. Sizing that only holds at full capacity holds during the hour nobody chose the unit for.
Stopping and starting repeatedly empties the sump by a different route. Copeland's application bulletin for its R-410A scroll compressors describes minimum run time as the span needed for oil thrown out at startup to find its way home and restore the level the crankshaft draws from. Cycling for less than that, the bulletin says, brings progressive oil loss and damage to the compressor. It also establishes that span on a system piped to the longest lines approved for it, which ties the two variables into one. Short cycling and a long route are the same shortfall arriving from opposite directions.
A charge below the specified weight thins the mechanism everywhere at once. Fewer kilograms of refrigerant crossing the same bore means a weaker stream doing identical carrying work, so an undercharge is a lubrication problem before it is ever a cooling one. That ordering is the part most people have backwards, and it is why topping up by feel is worse than it looks.
Idle branches on a shared condenser are the case nobody thinks to check. Copeland notes that circuits left inactive through part of the normal cycle can hold a meaningful quantity of oil, where the piping lets it fall out of the refrigerant flow into a branch nothing is passing through. A bedroom head switched off for a season is doing exactly that. The oil is not lost, it is parked, and the compressor is short by however much is parked.
| What changed | What the gas stops doing | Where the shortfall shows first |
|---|---|---|
| What changedSuction line one size too wide for the machine | What the gas stops doingTravels too slowly at the wall, where the film actually sits | Where the shortfall shows firstThe steepest part of the route, before anything flat suffers |
| What changedSystem settled at its minimum output | What the gas stops doingLoses the speed margin the line was sized around at full capacity | Where the shortfall shows firstThe tallest lift on the run, while level stretches carry on |
| What changedCompressor stopping before its run time is out | What the gas stops doingNever gets long enough to sweep home what each start threw out | Where the shortfall shows firstThe sump, falling a little further on every cycle |
| What changedRefrigerant weight under the figure the route calls for | What the gas stops doingMoves less mass through the same bore, so it pulls less along | Where the shortfall shows firstThe far end of the longest branch on the system |
| What changedOne indoor head left switched off for a season | What the gas stops doingPasses through that branch at all, so nothing sweeps it | Where the shortfall shows firstThe circuit used least, which is also the one nobody inspects |
Why does the same system lose a second compressor?
A compressor short of lubricant gives nothing away while it is happening. Bearing surfaces wear a fraction on every rotation that runs thin, the wear does not undo itself, and the room keeps cooling throughout. No remote reports it and no display carries a code for a film thinner than it should be.
The first outward sign points somewhere else entirely. Oil settling in the evaporator coats the inside of the tubing and gets in the way of heat transfer, which HVAC School names among the costs of a poor return. What that produces is a system cooling slightly less well than it used to. Weak cooling has a dozen likelier explanations, and every one of them gets examined first.
The delay finishes the job. The route was fixed on install day, the wear gathers across years, and the failure lands far enough downstream that nobody connects the two. The invoice says compressor. The explanation offered is age. Both are accurate, and neither is the cause.
A second failure on the same pipework is the signal, and it is the one worth acting on. The replacement goes back onto an identical route, at an identical size, with an identical climb and an identical charge. Whatever starved the first one has not gone anywhere. Age cannot account for a part that died young twice, and a system on its third compressor has stopped being a compressor story.
What that does not settle is whether the machine in place now is healthy. Symptoms belong elsewhere. The observable side of signs compressor failing is its own page, and what a compressor burnout leaves in the pipework is separate again. Electrical testing reads the compressor windings rather than the bearings, so a clean electrical result rules out less than it appears to.
The questions below get answered from the job rather than from the machine. That is deliberate. A route leaves a paper trail and a worn bearing does not. Whoever sized the run has figures ready. Whoever fitted what was to hand has an opinion, and the difference surfaces on the first answer.
- What suction line size this model calls for, and whether that figure came out of its installation manual.
- Whether the copper was reused from an earlier system, and what capacity that earlier system had.
- Whether the route includes a vertical rise, and whether a trap was fitted at the foot of it.
- Whether any long horizontal stretch sags or dips anywhere along its length.
- What weight of refrigerant went in, set against what the routed length calls for.
- How the unit behaves under light demand, whether it holds a steady low output or keeps stopping and restarting.
- Whether any indoor head on a shared condenser stays switched off for long stretches.
Ready to get started?
Tell us what’s going on. Symptoms, setup, photos, anything we should know. We’ll assess and come back with the right next step.