Cloudy Ice vs Clear Ice: The Six Causes and How to Test for Each
Cloudy ice comes from six causes: trapped dissolved air, high mineral load, a freeze rate that's too fast, no water agitation, biofilm in the water path, or a machine that was never built to make clear ice. Melting a cube in a clean glass separates the two big families in ten minutes — mineral cloudiness leaves a white ring, air cloudiness leaves nothing. On bullet and nugget machines the answer is usually the sixth cause, and no filter or water change will fix it.
Cloudy ice has six causes and, on most home machines, only two of them are worth an afternoon. Trapped dissolved air and dissolved minerals produce the great majority of white-cored cubes, and one ten-minute test with no tools tells them apart: melt a cube in a clean glass and see whether anything is left behind.
The other four causes — a freeze rate that's too fast for the water, a dead circulation pump, biofilm in the water path, and a machine that physically cannot make clear ice — each announce themselves with one specific observation. The table below pairs every cause with the measurement that identifies it and tells you honestly whether it's fixable on the machine you own.
Settle one thing before you spend money. If you have a bullet-style countertop machine or a nugget machine, cause six is almost certainly your answer, and no filter, no distilled water and no setting change will move it much. That's not a defect. It's how those evaporators work.
Two families of cloudiness that never overlap
Water freezes clear when the ice grows in one direction and slowly. A moving freeze front pushes dissolved gases and dissolved solids ahead of itself, so they end up concentrated in the last water to freeze — or flushed away entirely if that last water gets drained off. That's the whole trick behind directional-freeze clear cube makers, and behind every insulated-from-the-sides freezing method people improvise at home.
Freeze water from every side at once and there's nowhere for the gas and the minerals to go. They get trapped mid-cube as microscopic bubbles and crystals, which scatter light. White core, clear shell. That is family one: freeze-geometry cloudiness.
Family two is load. Even a perfectly directional freeze will look hazy if the water arrives carrying 350 ppm of dissolved solids, because there is simply too much material to push ahead of the front. Hard water also gives you a second tell — the cubes go brittle and crack, and the meltwater leaves a chalky film on the glass.
Six causes, six tests
Cause What identifies it (measurable) At-home test and expected result Fixable on that machine? Fix and realistic clarity gain Trapped dissolved air White core with a clear outer shell 2-4 mm thick; meltwater leaves no residue Melt one cube in a clean glass at room temperature. Water stays clear, glass dries with no film or grit Partly — depends on the evaporator geometry Use degassed water (boiled, then cooled sealed) and choose the machine's smaller ice setting. Core shrinks; it will not disappear on a bullet machine High TDS / mineral load Above roughly 250 ppm TDS, or hardness above 7 gpg Handheld TDS meter on tap water and on melted ice. Then evaporate a spoon of meltwater on dark glass — a white ring or grit confirms it Yes, at the water source only Carbon block does nothing for TDS. Reverse osmosis or distilled feed drops it under 50 ppm; the step change shows in the second or third batch Freeze rate too fast Cycle length shorter than that class's normal: under 6 min on a bullet unit, under 14 min on a clear-cube unit Stopwatch three consecutive cycles. Compare against 7-13 min bullet, 15-25 min clear-cube, continuous extrusion on nugget Only on machines with an ice-size or thickness selector Pick the larger or thicker setting so the front advances slower. Modest gain — expect a smaller core, not a clear cube No agitation (circulation pump failed) Machine used to make clear ice and no longer does; no water movement audible or visible over the evaporator during freeze Open the lid mid-freeze on a clear-cube unit. Water should be sheeting visibly over the plate. Still water means the pump is down Yes — the pump is a serviceable part Replace the circulation pump, roughly $12-30 countertop and $40-90 undercounter. Clarity returns fully if the pump was the only fault Biofilm contamination Slimy or pink-orange film on the reservoir wall or water path; cloudy ice with a musty or sweet-plastic smell Wipe the reservoir wall with a white paper towel. Colour or slime transfers onto the towel Yes, but it returns without a schedule Full descale, then sanitise, then fix an interval. Clarity and smell both recover; skip the interval and it's back in 4-8 weeks Machine-class limitation Bullet, nugget or flake machine — opaque ice is the design output None required. Nugget is compacted flake and holds air by construction; bullet freezes outside-in onto a chilled prong with the core last No No fix exists. Clear ice needs a directional-freeze evaporator with flowing water over it. Change class or accept the ice Work top to bottom and stop at the first row whose test comes back positive — the causes below it can't be diagnosed cleanly until the one above is cleared. Two rows are worth a second look. The agitation row is the only one that explains a sudden change: pumps fail, and a machine that made clear ice last month and cloudy ice this month is far more likely to have a dead pump than a changed water supply. The machine-class row is the one that saves the most money, because it stops people buying RO systems for a machine that would still make white ice on distilled water.
Order the tests so each one eliminates the most
Melt test first. It costs nothing, needs no instrument, and it splits the six causes into two groups of three in about ten minutes. Residue on the glass means you're in the mineral group. Clean glass means air, freeze rate or machine class.
TDS meter second, if the melt test left residue. A handheld pen meter runs $12-20 and is the single most useful instrument in this category after a multimeter. Meter the tap, meter the reservoir, meter the melted ice. Ice always reads lower than its source water because the freeze front rejects some of the solids — if melted ice reads within 10 ppm of the tap, the freeze is going so fast that nothing is being rejected, and you've just confirmed the freeze-rate row too.
Cycle timing third. Everything else in this category degrades slowly, so a stopwatch on three cycles gives you both a clarity answer and a baseline you'll want later when output starts dropping.
- Melt one cube in a clean glass, room temperature, undisturbed — check the glass after it dries
- TDS pen on tap water, reservoir water and melted ice, recorded in ppm
- Stopwatch on three consecutive complete cycles, recorded in minutes
- White paper towel wiped along the reservoir wall and around the pump inlet
- Lid open for ten seconds during a freeze cycle to confirm water is moving
What your water is actually carrying
TDS (ppm) Hardness (gpg) What the ice looks like Scale risk inside the machine Action that changes it 0-50 Under 1 Clear apart from any trapped-air core Negligible; descale twice a year for hygiene, not scale Nothing — you're already at the floor 50-150 1-5 Faint haze in the core only Low; descale every 10-12 weeks Carbon filter for taste; it won't change clarity 150-250 5-10 Visible white haze through most of the cube Moderate; descale every 8 weeks or cycles start lengthening Polyphosphate cartridge slows scale; RO is what fixes clarity 250-400 10-17 Chalky white throughout; meltwater rings the glass High; evaporator film builds in 4-6 weeks and drags output down RO on the ice maker feed. Nothing cheaper holds 400-600 17-25 White, brittle, cubes crack and weld together in the bin Severe; measurable output loss within a month of a descale RO, and shorten the descale interval to 4 weeks regardless Over 600 Over 25 Opaque, soft-edged, often with visible flecks The machine will not hold a healthy Harvest Index at all RO is mandatory; untreated feed at this level shortens machine life by years Meter your own supply rather than trusting the municipal average — softener bypasses, well water and long service lines move a house well off the published number for its zip code. Hardness in grains per gallon is just TDS expressed differently for the calcium and magnesium fraction: divide ppm of calcium carbonate by 17.1. A TDS pen reads everything dissolved, including sodium from a softener, which is why softened water can read high on TDS while causing almost no scale. Softened water also makes noticeably cloudier ice than RO water at the same TDS reading, which surprises people who just spent money on a softener.
Clear ice is an evaporator design, not a setting
Three evaporator types dominate home machines and each one determines the clarity ceiling before water ever enters it.
Bullet and tube machines dip chilled metal fingers into a water bath. Ice forms on the outside of the finger and grows inward, so the last water to freeze is at the centre of the cube with nowhere to reject anything to. That trapped centre is the hollow-looking white core people describe, and it's structural.
Nugget machines freeze a thin film on a chilled cylinder wall, then scrape it off with an auger and compress the flakes into pellets. The ice is opaque because it is thousands of small crystals with air between them. That air is exactly what makes nugget ice chewable and absorbent, so the cloudiness and the reason people want the ice are the same property.
Directional-freeze cube makers spray or sheet water over a cold plate so that ice builds up in layers with moving water carrying gas and solids away. That's the only common home design that can make genuinely clear cubes, and it pays for it in cycle length and in purged water.
It used to be clear. Now it isn't.
A change is a different problem from a limitation, and it has a much shorter suspect list. Four things cause a machine that made acceptable ice to start making bad ice.
The circulation pump dies or its inlet screen blocks with scale. Water stops moving over the plate, the freeze stops being directional, clarity collapses in one cycle. This is the most common single cause of sudden change on clear-cube machines and it's cheap to fix.
Scale films the evaporator. A mineral layer insulates the plate, so the machine runs longer per batch and freezes less uniformly — you get cloudy ice and a falling Harvest Index in the same week. If your cycles have lengthened at the same time clarity dropped, descale before you buy anything.
Biofilm establishes in the reservoir or tubing. Usually arrives with a smell and often with a slimy feel on the reservoir wall. It clouds ice by seeding crystal formation everywhere at once.
Or the water supply changed under you — a new filter cartridge that isn't the same media, a softener regenerating on a different cycle, seasonal source switching by the utility. Meter the tap before assuming the machine did something.
What clarity you can realistically buy back
Be blunt with yourself about the ceiling. On a bullet machine, degassed water plus the small-ice setting gets you from an obvious white core to a smaller white core. That's it. Owners who chase clarity on that class usually end up disappointed and out the cost of a filter.
On a clear-cube countertop or undercounter machine, the same interventions matter enormously, because the design can actually use them. RO water at under 50 ppm through a healthy circulation pump on a descaled plate produces cubes that are clear except for a small centre line — which is the honest best case for a home machine, and is better ice than most bars serve.
Nugget owners should skip this entire exercise unless the ice smells or tastes wrong. Cloudy nugget ice is working nugget ice. If it smells, you have a biofilm problem, and that's a cleaning-schedule question rather than a clarity question.