Suction accumulator: the trap ahead of the compressor

A suction accumulator sits on the pipe between the evaporator outlet and the compressor inlet. Henry Technologies, which builds them, gives the primary function as preventing a sudden surge of liquid refrigerant, or oil, from returning down the suction line and into a compressor. Its own words for the part are a temporary reservoir for liquid refrigerant and oil.

Everything about the arrangement follows from one property of a compressor. It moves gas by shrinking the space that gas occupies, and a liquid has almost no space to give up. Danfoss states the consequence plainly in its application guide for scroll compressors: floodback can cause oil dilution and, in extreme situations, lead to liquid slugging that damages the compressor.

Sitting on the suction side is what makes the vessel useful and also what makes it awkward. Everything the evaporator sends back has to pass through it, including the oil the compressor needs returned. A vessel that catches liquid catches oil with it, and that is the design problem the rest of this page keeps circling.

The part takes two quite different physical forms. On designed plant it is a freestanding steel shell plumbed into the suction line, selected and sized by whoever engineered the system. On room-sized equipment it is a small canister fixed to the compressor housing, supplied with the compressor and never handled on its own.

Two separate kinds of damage sit behind the design, and only one is dramatic. A surge of liquid can hydraulically lock a compressor in a moment. A slow return does something quieter. Copeland's application guidelines warn that liquid refrigerant dilutes the oil, can wash the oil off the bearings and moving parts, and can lead to local overheating and compressor failure. The second path leaves no noise on the day it starts.

Nothing about it is visible from inside a flat. It has no moving parts, no control wiring and no sensor, so it produces no reading and no error code. What it does happens in the seconds after the coil sends back something the refrigerant circuit was never supposed to produce.

The circuit is designed so that no liquid ever reaches the suction line, which is exactly why the vessel exists. Refrigerant is metered into the evaporator, boils across the coil, and leaves as vapour carrying a margin of extra heat. Danfoss sets that margin as a suction superheat between 5 and 30 K in any conditions, held by the expansion device.

Floodback is the name for that assumption breaking. Danfoss defines it as part of the refrigerant entering the compressor still being in a liquid state. The definition is worth holding onto, because it covers two events that behave nothing alike. A steady trickle dilutes the oil over months. A sudden surge arrives all at once.

Too little heat at the coil produces the steady version in cooling. Danfoss runs its floodback test at a high pressure ratio and minimum evaporator load. That is the corner of the map where the coil cannot finish the boil. Copeland reaches the same place from the other side. It names a partly blocked indoor air filter, or a loss of evaporator airflow, as the test to run on a cooling design.

A metering device that has stopped metering does it more directly. Copeland singles out orifice refrigerant control. Large volumes of liquid flood back there during normal steady running, and the oil is diluted until the bearings run short of it. A thermostatic valve holds superheat instead. Copeland allows that a heat pump built around one may need no accumulator, if testing proves the point across the range.

Swings in load count separately from steady running. Danfoss asks for a transient test under fan staging, compressor staging and similar changes, and adds a defrost test for reversible machines. Copeland treats the start and end of defrost as the moment to assess liquid slugging and oil pump-out. A machine restarted after a long stop has a different problem again, and refrigerant migration covers that one.

Vapour leaves from the top and liquid stays at the bottom, which is the whole idea in one line. Henry Technologies describes vapour return through a special U tube arrangement, with a tube within a tube used on certain models. Liquid is held at the bottom of the shell, ready for metering back to the compressor.

Oil is what turns a simple separator into a metering device. Whatever the coil sends back, oil comes with it, and oil sitting in the shell is oil the compressor is no longer running on. Henry's answer is a screened orifice at the bottom of the tube. Liquid and oil are fed back through it at a controlled rate while the vapour carries them along, and that metering happens only while the compressor runs.

The hole doing that work is small, and its size is settled at design and not in the field. Copeland's R-410A scroll bulletin puts the accumulator oil return orifice at 0.040 to 0.055 inches, or 1 to 1.4 mm. The choice turns on compressor size and on measured floodback. A hole that size lives or dies by what reaches it.

The screen guarding it is deliberately coarse for the same reason. Copeland requires a large-area protective screen no finer than 30 by 30 mesh, which is 0.6 mm openings, and warns that finer screens plug with normal system debris and starve the compressor bearings of oil. The same bulletin extends that limit to every screen in the system.

Fitting the vessel also changes the oil balance of the whole machine. Copeland's guidelines tell the designer to ask the supplier whether the accumulator needs an extra charge of oil. The system is then pre-charged to suit. Danfoss makes the same point from the flow side. Its guide notes that the vessel's size affects gas velocity and oil return hole size, so oil return has to be checked again.

Debris does not have to come from a failure to end up at that orifice. Copeland's brazing caution describes copper oxide forming inside pipe joined without a nitrogen flow, then being swept through the system to block the screens protecting expansion valves, capillary tubes and accumulator oil return orifices. Work done years earlier can end up sitting in front of a 1 mm hole.

Judging any of this from outside the shell is not possible, and the makers say so. Copeland asks that the accumulator's ability to prevent liquid slugging and oil pump-out be assessed during system development, and notes what that takes: special accumulators and compressors fitted with sight tubes or sight glasses, so refrigerant and oil levels can actually be watched. A finished vessel bolted into a machine gives up nothing.

Compressor makers treat the accumulator as a concession, not a default fitting. Copeland says its scroll compressors handle liquid during occasional floodback well enough. On that basis it calls an accumulator unnecessary in standard designs such as condensing units. Its European guidelines go further, and state that one is not needed for durability in most systems. The vessel earns a place only where the design cannot promise dry vapour.

Heat pumps are the standard case for fitting one, and metering is the reason. Copeland requires an accumulator where orifice control lets liquid flood back during normal steady running. It then allows that a heat pump metered by a thermostatic valve may need none. That exemption rests on testing across the operating range. The vessel stands in for control the design does not otherwise have.

Danfoss lists the accumulator last among its remedies, and the order carries the lesson. Where a floodback test fails, the first moves are the expansion valve's selection and setting, the bulb position on a thermostatic valve, the sensor chain and control loop on an electronic one, and the defrost logic on a reversible machine. Adding a vessel comes after all of that. A machine that needs one to survive normal running has a metering fault the vessel is covering up.

Not every vessel hanging off a vapour line is an accumulator, and the confusion is easy to fall into. Copeland describes a charge compensator as a hollow shell with a single opening, tapped into the vapour line to park surplus refrigerant while a valve-controlled heat pump runs in heating. It holds charge the coil cannot use. An accumulator holds liquid the compressor must not receive. Same pipe, same look, opposite jobs.

Size follows the worst case a system can produce, and the published figures sit far apart because those worst cases do. Two compressor makers show the spread on the same component.

None of those numbers transfer to a room air conditioner. They describe vessels an engineer selects while a system is still on paper, for duty a Singapore flat never asks for. Equipment sold for a bedroom arrives with its own answer already built in.

Wall splits here run rolling-piston rotary compressors. That family carries its accumulator on the compressor itself. CIBSE's technical module on the type describes a suction accumulator attached to the compressor housing. It works as a trapping device, so that returning liquid refrigerant is boiled off before it enters the compressor, along with the oil coming back with it. The module calls it a further safeguard on a design that already copes fairly well with slugging.

Service manuals draw it, then decline to sell it separately. Mitsubishi Electric's manual for its cooling-only split-type air conditioners shows an accumulator in the refrigerant system diagram. It sits on the suction side, ahead of the compressor, beside the muffler and the capillary tube. The outdoor unit parts lists in the same manual name the compressor. They do not name an accumulator beside it.

So the honest answer for a household unit is yes, invisibly, with nothing to inspect. It will not appear on a quote. No technician will offer to change it on its own. If a wall split is ever said to need an accumulator fitted, ask what measurement showed liquid arriving. That request describes a design change, not a repair.

What deserves the attention is the thing the vessel quietly covers for. Liquid reaching a compressor is a downstream symptom, and the causes above it are the ones a diagnosis can act on. Airflow starved at the indoor coil, a charge that does not match the pipe run, a metering device no longer holding superheat. Those are repair work. The canister behind them is not.

One rule from the primaries does travel into real service work. Copeland says that after a motor burn, a system containing an accumulator should have it replaced, because the orifice or screen may already be plugged with debris or may plug shortly afterwards. The outcome it names is oil starvation to the replacement compressor and a second failure. Where compressor burnout has contaminated a circuit, the cheapest part on the job decides whether the expensive one survives.