Ice Maker Makes Hollow Cubes Decision Tree
Why this matters
Hollow or partially-filled ice cubes are a specific symptom that almost always points to insufficient water in the mold during freeze - and the root cause is almost always upstream of the ice maker itself. Techs who default to "bad ice maker module" replace a $90 part on a $30 inlet-valve problem and create an unnecessary callback when the new module produces the same hollow cubes. The decision tree below works from water supply outward to the maker, catches the cheap fixes first, and reserves head/module replacement for actual head failures.
Symptom presentation
Customer reports ice cubes that look like shells - the outer wall freezes but the center cavity is empty or filled with water-slush. They harvest, fall into the bin, and break easily. Often paired with reduced production volume; sometimes paired with a wet ice bin from cubes that were ejected before fully freezing.
Two distinct mold designs in the field:
- Old-style "tin can" molds (Whirlpool/Maytag/KitchenAid heritage design): 8 crescent-shaped cavities in a rotating mold with a thermostatic harvest cycle.
- Twist-tray molds (Samsung, LG, Frigidaire newer platforms): plastic tray that twists to eject cubes after freeze sensing.
Both designs depend on a fill cycle dropping a precise volume of water into the mold; hollow cubes mean the fill volume was short.
Quick checks (under 10 minutes)
- Saddle valve at the supply line. Saddle valves are notorious for clogging with mineral or shipping debris and never fully open. Per UPC 605.5 saddle valves are limited or prohibited in some jurisdictions for that reason. If the install used a saddle valve, recommend replacement with a 1/4-turn ball valve regardless of the immediate diagnosis.
- Water filter age. Filters older than 6 months reduce flow significantly. Pull and bypass to confirm.
- Incoming water pressure. Spec is 20-120 PSI per most OEMs (e.g., Whirlpool W10238110 spec). Below 20 PSI starves the inlet valve. Test with a hose-bibb gauge at the saddle or supply takeoff.
- Inlet valve fill time during a cycle. Force a harvest with the OEM service-mode jumper or by turning the bail wire manually. Time the fill cycle audibly. Whirlpool old-style targets 7.5 seconds at 20 PSI for 110-140 mL fill volume. Substantially shorter = inlet valve flow restriction.
- Freezer temperature. Above 10F slows the freeze cycle; the harvest thermostat may eject before the mold is fully frozen. Spec is 0F to -5F at the ice maker mounting bracket.
Isolation tree
Branch 1 - Restricted water supply (~35%). Saddle valve partially closed, mineral-clogged supply line, kinked plastic supply tubing behind the unit. Most common cause overall. Cycle fill time will be visibly long with reduced flow. Replace saddle valve with 1/4-turn shutoff and inspect tubing.
Branch 2 - Clogged water filter (~20%). Filter past service interval. Bypass the filter (most units have a bypass plug) and run a forced harvest. Fill time returns to spec = filter is the cause. Replace.
Branch 3 - Inlet valve diaphragm or solenoid (~15%). Inlet valve passes water but flow is reduced from internal restriction. Ohm the solenoid (typical 250-500 ohms cold), inspect the inlet screen for debris. Replace the valve - do not attempt to clean and reinstall, the diaphragm wears with debris.
Branch 4 - Low household pressure (~10%). Well systems with marginal pressure tank charge, or municipal supply during peak demand. Pressure gauge at the saddle valve shows the truth. Solution is plumbing-side, not appliance-side; document and refer.
Branch 5 - Twist-tray sensor or thermistor (~10%, twist-tray models). Samsung and LG twist-tray makers use a thermistor on the tray to detect freeze completion. Drifted thermistor ejects before the freeze finishes. Service mode shows tray temp; compare to OEM table.
Branch 6 - Faulty ice maker head / module (~5%). Old-style modular head: harvest motor turns the mold to eject before the fill cycle has fully run, or the head logic skips the fill step entirely. Confirm only after upstream water supply is verified. Replace as a complete module.
Branch 7 - Ambient temp too high (built-in beverage cooler models, ~3%). Some standalone ice makers struggle at high ambient. Spec is typically 55-90F ambient; outside that range performance degrades by design.
Branch 8 - Freezer temperature too high (~2%). Door seal, defrost system, or sealed-system fault upstream. Hollow cubes are a symptom; the freezer fault is the real call.
Confirming diagnosis
- Supply restriction: measured pressure at the saddle and observed fill volume. Catch the fill in a graduated cup during a forced harvest - 110-140 mL at 7.5 seconds is healthy for old-style modular. Below 80 mL = restriction confirmed.
- Filter: before/after bypass test. Volume returns with bypass = filter confirmed.
- Inlet valve: ohm solenoid and inspect inlet screen with the valve disconnected. Visible debris in the screen with low flow even after screen cleaning = valve replacement.
- Pressure: gauge reading at supply takeoff. Static pressure plus dynamic pressure during fill.
- Thermistor (twist-tray): OEM service mode readout or ohm at room temperature against the OEM table.
- Head/module: substitution test. Install a new head, run a forced harvest, time the fill cycle. If a new head fixes a fill problem when supply is verified, head was the cause (rare).
Remediation - repair vs replace ROI
| Fault | Repair viable when... | Replace ice maker / refrigerator when... |
|---|---|---|
| Saddle valve | Always - replace with 1/4-turn ball valve. Plumbing work, not appliance. | Never as an appliance call. |
| Water filter | Always - $20-50 OEM filter, 5 minutes. | Never. |
| Inlet valve | Always. MSRP $40-90, 30 minutes. | Never. |
| Low household pressure | Refer to plumbing. | Never as an appliance call. |
| Thermistor (twist-tray) | Under 8 years. MSRP $20-50. | Over 10 years on a Samsung twin-cooling platform where thermistor access requires sealed-system disturbance. |
| Ice maker head / module | Under 12 years. MSRP $80-180. | Over 15 years OR head plus inlet valve plus filter housing all marginal. |
| High freezer temp from upstream fault | Address the upstream sealed-system or defrost fault. | Refrigerator at end-of-life on age. |
Cutoff guidance: hollow cubes calls are almost always a saddle valve or filter problem and almost never an ice maker module problem. Bill the diagnostic, fix the upstream cause, and document the operator-side maintenance schedule (filter every 6 months minimum, saddle-valve-to-ball-valve upgrade strongly recommended). Customers who decline the upstream fix get a "deferred maintenance" disclaimer on the invoice - the call will recur within 6 months and you want the documentation.
References
- UPC 605.5 (saddle valve restrictions per Uniform Plumbing Code)
- UL 250 Standard for Household Refrigerators and Freezers (ice maker safety requirements)
- Whirlpool Service Pointer W10238110 (ice maker fill volume and harvest cycle timing)
- Samsung Service Bulletin (twist-tray ice maker thermistor diagnostic, RF series)
- DOE 10 CFR 430 Subpart B Appendix A1 (refrigerator energy test including ice maker on cycle)