Hollow and Fused Cubes From a Cambridge Sub-Zero Ice Maker, Diagnosed

Reach into the bin and the evidence is already in your hand. Some weeks it is a fistful of thin shells that crush between your fingers. Other weeks it is one frozen plate you have to break loose with the scoop before anything else will come out. Those two handfuls describe opposite problems, and nothing on the control panel will tell you which one you have.

An ice maker does two jobs that are easy to confuse. It fills a mould with water and freezes it, and then it lets go — warming the mould just enough to break the bond and pushing the batch out into the bin below. Nearly every complaint that reaches us in Cambridge is one of those halves misbehaving, and the half at fault is written into the shape of what came out.

So the useful order of work is this: read the ice, follow the water, find out how cold it really is where the mould sits, and only then start asking whether the release side is running late.

Read the cube before you touch the appliance

Four things are worth noticing before anything gets dismantled: shape, clarity, size and whether the pieces are stuck to each other. A dished or hollow cube says less water arrived than the cycle expected. Cloudiness comes from dissolved air and minerals leaving solution as the ice freezes. Small but well-formed pieces are a third, separate signal. A welded slab is a fourth.

Hollow is the loudest of the four. Freezing runs inward from the walls of the mould, so a short fill leaves a shell with an unfrozen middle that collapses when you pick it up. Cloudy, on its own, is cosmetic and turns up on machines with nothing wrong with them at all. Small-but-neat means each fill is slightly shy or the freeze is being cut short — the mechanism is fine, the arithmetic is off.

The slab is the one people misread most. Two households in East Cambridge with the same appliance will get different ice out of it: the one that empties the bin every evening rarely sees a slab, the one that entertains twice a year almost always does. Ice left standing eventually welds to itself, and that is a household pattern rather than a fault.

Supply line, valve and pressure

If the pieces are hollow or short, start on the water side, and start upstream of the appliance. A stop valve somebody closed, a line pinched behind the cabinet, or an old saddle fitting that never opened all the way will each starve the fill without ever shutting it off outright — which is precisely why the symptom looks like a failing machine.

Kitchens around Harvard Square are frequently carved out of former pantries and back halls, and the stop for the water line ends up behind a cabinet run nobody has opened since the work was signed off. A plumber closing valves to get at a bathroom stack, somebody shutting things down during a leak scare — the ice thins out a day later and the two events never get connected.

A crushed line is the other common one. When a full-height built-in in a Porter Square triple-decker is pushed back hard after a floor is refinished, the soft copper behind it takes the load. The fill is timed rather than measured: the valve opens for a set period, and whatever the pressure behind it can push through in that window is what the mould receives. Lower the pressure and you get the same cycle carrying less water, batch after batch.

When was it last changed?

Ask about the filter before anything comes apart. A loaded cartridge restricts flow well before it does anything to taste, so water that still tastes clean tells you nothing about the state of the filter. Ice that has quietly got smaller over a couple of months is a filter story more often than it is anything else.

Aftermarket cartridges belong in the same conversation. Fitting the housing and sealing properly is not the same as passing the flow the fill cycle was designed around, and a cartridge that does everything else right can still meter the water. Running the unit in bypass is a fair test: if the pieces come back to full size, the filter was the restriction. If they do not, it was not — and that is the whole of what bypass proves. It says nothing about the valve or the line.

Intervals vary by series and by how hard a household draws on it, so use the figure printed in the use and care guide for the model you actually own rather than one a neighbour quotes you.

When the fault is not in the ice maker at all

Ice is made at the mould, not at the display. A compartment reading its set point on the panel can still be running warm at the spot where the mould sits, and a mould that never gets down to temperature on schedule turns out small, slow, badly formed pieces with nothing whatsoever wrong with the ice maker itself.

Loading is the usual culprit. Bags and boxes stacked against a return vent stop air moving where it has to move, and the space around the mould is one of the first places to feel it. Anything packed away while still warm makes the next several hours worse.

Then there is the door. Production is rated over a full 24 hours, and that rating quietly assumes a door that mostly stays shut. An evening of guests helping themselves to drinks knocks recovery back at exactly the moment demand peaks — which is why the bin empties during a party and fills itself again by Sunday lunchtime with no repair in between.

Cubes that form and never leave the mould

Harvest is the release half of the cycle. A small heater warms the mould just enough to break the grip, and an ejector sweeps the batch out into the bin. When harvest runs late, runs weak, or has to try twice, the ice is already made — it simply stays where it is, or lands in the bin still stuck together.

That is why a fused slab points at timing, or at low turnover in the bin, rather than at freeze capacity. The machine has already proved it can freeze. The open question is when it let go.

Two behaviours are worth watching without touching anything. Does the ejector move at all, and does the sensing arm rise and fall freely? An arm propped up by a slab reads as a full bin and production stops completely, so a jam presents to the household as a dead ice maker. Stop there. Do not chip at the mould with anything metal and do not force the arm back down; both are how a small fault turns into a parts order. Our Sub-Zero Ice Maker Repair page sets out what that visit covers.

Handing it over

Water on the floor, a line dripping behind the cabinet, or a compartment that will not hold temperature are calls rather than experiments. So is a machine that has gone completely silent with a slab jammed against the arm.

A visit establishes the part you cannot get at from the kitchen: how much water actually enters the mould, what the inlet pressure does while the valve is open, how long a cycle runs from fill through to harvest, and the temperature at the mould rather than the number on the display. Those four together separate a supply fault from a release fault in one sitting.

Access is worth raising when you book. A Kendall Square tower means a loading dock and a lift that has to be reserved, and a built-in that went into a Porter Square kitchen during a gut renovation is not coming forward on its own. We work across Kendall Square, Harvard Square and Porter Square. Ring the number at the head of this page, or use the booking link beside it; those are the two routes to us.

Still asked on the phone

Are hollow cubes always a water supply problem?

Not always, but that is the first place to look. A hollow cube means less water entered the mould than the cycle expected, and the usual reasons are a partly closed stop, a crushed line or a saturated filter. The same shape can also come from a freeze that is being cut short, so if the water side checks out clean, the next question is how cold it really is where the mould sits.

Does an aftermarket filter cartridge matter?

It can. Fitting the housing and sealing correctly is not the same as passing water at the rate the fill cycle was designed around, and a cartridge that meters the flow produces the same short fills as a half-closed valve. If the ice recovered after a change and then shrank again sooner than it should have, rule the cartridge out by running the unit in bypass.

Should I throw away the first batch after a filter change?

Yes, discard the first few batches. Changing a cartridge puts air into the line, and air in the fill shows up as cloudy, uneven and sometimes undersized pieces for a cycle or two while it clears. Judging a new filter on the first batch tells you almost nothing. Give it a day of ordinary use and then look at what the bin is actually producing.

ice makerwater supplydiagnostics