Aircon crankcase heater: why the oil has to stay warm

Refrigerant in a stopped system does not hold position. It moves towards whichever part of the circuit is coolest, and on an outdoor unit sitting overnight that is often the compressor itself. The steel shell is heavy and sheds heat slowly. By early morning it can be the coolest large mass in the loop.

Temperature is only half of what pulls the gas there. Compressor oil and refrigerant mix readily, so oil in the sump takes refrigerant vapour in rather than pushing it away. Taking it in drops the pressure just above the oil surface, and lower pressure draws more gas across. The process feeds itself, quietly, for as long as the machine stays off.

What gathers is not a film lying on top. It goes into solution, so the sump ends up holding something that looks like oil and behaves like a thinner fluid. Oil carrying dissolved gas is less viscous than the oil the bearings were designed around. That counts for little while the machine is off, and for a great deal in the first seconds after it starts.

Standstill is the condition that produces all of this, so the behaviour sits with stopped machines rather than faulty ones. Nothing has gone wrong. A healthy system parked for a week does exactly this, and the pipework gives no outward sign. A pump down answers the same physics from the other end, by cornering the charge behind closed valves so there is nothing left in the rest of the circuit to move. That is a service and relocation step, not something a machine does for itself every night.

A crankcase heater has one job, and it is a temperature job rather than a heating job. It holds the oil in the compressor warmer than the rest of the circuit. Gas gathers at the coldest point available. If the sump is no longer that point, there is no reason for anything to settle in it.

Warmth also works on what has already gone into solution. Refrigerant comes back out as the oil temperature rises, so a warm sump lets it go gradually across the hours the machine sits idle. Gradually is the word carrying the weight. The same gas leaving in a rush is the event the whole arrangement exists to prevent.

The rush arrives at start-up. Pressure in the crankcase falls the moment the compressor pulls down, and gas that sat comfortably in solution at the old pressure boils out at the new one. The oil turns to foam. Foam compresses, carries almost no load, and strips off the bearing surfaces instead of coating them. For those first seconds the pump is moving froth, and the moving parts run on very little.

Two further outcomes ride on the same event. Foaming drives oil out of the sump and up the discharge line, so what remains can finish low as well as thin, and oil that has left has to find its way back through the pipework. Where liquid reaches the compression chamber, nothing in that chamber was built to squeeze a liquid. Both outcomes are mechanical, and neither raises an error code while it is happening.

The electrical half of the compressor sits outside all of this. Compressor windings fail for their own reasons and get measured their own way, and a bearing worn by dry starts leaves no trace on a winding reading. Keep that separation in view. It explains why a clean electrical test settles less than people expect.

Fitment follows the quantity of refrigerant, not the type of building. Manufacturers write the requirement as a threshold, and the threshold is about how much liquid is available to gather in the sump. A system holding a little has little to lose. A system holding a lot has a real problem waiting for it.

Copeland states the rule as a charge limit for its scroll compressors. Its technical information for the range sets a limit per compressor family, running from half a kilogram at the small end to over thirteen kilograms at the large. A heater is required once the system charge passes whichever limit applies. The document adds that this does not depend on the type or layout of the system, and that compressors running carbon dioxide always need one.

Carrier writes the same idea as a pipe length. Its C4H5S installation instructions call for a heater where the refrigerant tubing runs longer than 80 ft, which is 24.38 m. Length and charge are one variable read two ways, because a longer route holds more gas.

That is why the subject lands on commercial plant more often than on a household split. Charge climbs with the number of indoor units and with the distance to the condenser, so the machines most likely to sit above a threshold are the large ones. Whether a particular system is above its own threshold is a paperwork question. The manual answers it and the installer answers it. Looking at the casing does not.

Take the supply away for a long stretch and migration runs unopposed on any machine that has a heater. How much gathers tracks how long the stop lasted and how cool the surroundings stayed. A week and a night are not the same case, and the week is what this page is about.

What a parked system draws from the supply is a separate question. Whether cutting that supply off earns anything is another. Both sit with standby power, and both are answered there.

Damage is dealt in the first seconds of the restart, and it accumulates rather than lands all at once. One dry start does not destroy a compressor. Bearing surfaces wear a little each time they turn without a film, and that wear does not undo itself. A machine restarted carelessly after several long stoppages fails months later, by which point nobody connects it back to a shutdown that felt uneventful.

A competent restart is short to describe and slow to carry out. The system has to be live, with its heater doing its work, for the period the manufacturer names before anything asks the compressor to turn. That period is an instruction from the machine's own paperwork. The numbers behind it differ from one machine to the next.

Three published figures show how far apart they sit. Copeland's technical information for its scroll compressors sets 12 hours ahead of starting the compressor, under a warning about oil dilution and bearing malfunction. The Daikin AGZ070EH installation, operation and maintenance manual asks for 8 hours before the compressors are started. Carrier's C4H5S installation instructions require a minimum of 24 hours. Read the machine you actually own, not one of those three.

Nothing here tells anyone to touch a switch. Whether a system holds a heater, whether that heater still works, and how long it needs, are questions for someone who knows what is inside the casing. The owner decides when the machine comes back into use. Carrying that out is a separate job.

Useful preparation happens before the machine goes dark, not afterwards. Once the supply has gone the questions stay the same, but the answers start costing a visit. A few minutes with the manual and the person who installed the system covers all of them.

A flat left standing empty is the hardest version of the case. The stop is long by definition, and nobody is watching the machine through it. What stillness and Singapore humidity do to an aircon left empty is treated separately. The crankcase question is an addition to that list, not a substitute for it.

Commercial premises carry more of the risk and more of the warning. Larger plant is likelier to hold a heater, stoppages tend to be planned rather than accidental, and a restart usually has a date on it. Fitting a manufacturer's waiting period into that plan is a scheduling matter, and business aircon servicing already runs to a schedule.

Symptoms afterwards make a poor guide, which is precisely why this gets settled in advance. A compressor wearing from repeated dry starts says nothing at the time. What it eventually does say is covered under signs compressor failing, and by then the shutdown behind it is months in the past.

Compressor oil is a bigger topic than the part that warms it. What it is made of, how it breaks down, and why the wrong grade causes trouble all sit outside this page.

• •Whether this machine holds a heater at all, which the manual states and the installer will know.
• •The waiting period named in that machine's manual before the compressor next turns.
• •How long the stoppage will genuinely run, counted in days rather than in nights.
• •Who is accountable for the system being live again in time, and whether they will be on site to confirm it.
• •Whether the machine has already sat dark for a long spell, because that changes what the restart needs.