Industrial Chiller Troubleshooting: Common Faults, Causes, and Fixes

Key takeaways

• Most "chiller not cooling" complaints trace back to heat rejection (dirty condenser, failed fan, hot machine room) or flow (clogged strainer, tired pump), not the refrigerant circuit.
• Troubleshoot in order: confirm the symptom with a thermometer and flow reading, check the cheap and common causes first, and only then suspect refrigerant or controls.
• Freeze alarms are usually a setpoint or flow problem. Verify both before swapping the sensor.
• Refrigerant work belongs to certified technicians. Coils, strainers, fluid, and sensors are fair game for in-house maintenance.
• Log every chiller fault as a downtime event with a cause code so chronic problems show up as data, not anecdotes.

Process chillers sit quietly behind spindles, laser sources, injection molds, welders, and hydraulic power units, and when one faults, the machine it cools stops with it. This guide walks maintenance technicians and plant engineers through the common chiller faults, their most likely causes in order, and the first check to run for each.

Start here: symptom, likely cause, first check

Before opening anything, confirm what the chiller is actually doing: read supply and return temperature, note any alarm code, and look at the flow or pressure gauge. Then use this table to pick your starting point.

High temperature or not cooling

Work the heat rejection side first, because that is where most of these faults live. A dirty condenser coil or blocked air filter is the single most common cause: the chiller cannot dump heat, head pressure climbs, and capacity falls. Look at the coil with a light, and clean it with a fin comb, vacuum, or low pressure air blown against the direction of normal airflow.

1. Condenser coil and filters: clean and confirm airflow is unobstructed on both faces.
2. Condenser fan: confirm it runs at full speed and spins the right direction after any motor or wiring work.
3. Machine room ambient: chillers rated for 35 degrees C ambient lose capacity fast in a 45 degree corner where they share air with compressors and hydraulic units. The same heat rejection logic applies there too; see our guide to air compressor troubleshooting.
4. Fouled evaporator or heat exchanger: scale and biofilm insulate the plates, so the refrigerant side gets colder while the process fluid stays warm. A widening gap between refrigerant suction temperature and fluid supply temperature is the classic sign.
5. Low refrigerant from a leak: oil staining around fittings is the visual clue. Diagnosis and recharge are strictly work for a certified technician (F-Gas in the EU, EPA Section 608 in the US).
6. Undersized for the load: if the fault appeared after a new laser source, an extra mold circuit, or a duty increase, measure the real heat load before blaming the chiller.

A quick worked check: measure supply and return fluid temperatures at full load and compare the delta to the design value on the nameplate or datasheet. A larger than design delta with normal flow points to overload; a normal delta with a high absolute temperature points to heat rejection.

Low flow and flow alarms

Flow faults starve the process and can freeze the evaporator, so chillers alarm early on them. Check in this order:

• Strainer or filter: the number one cause. Isolate the circuit, pull the basket, and inspect. If it clogs repeatedly, the debris source (corrosion, algae, pipe scale) is the real problem.
• Circulation pump: listen for cavitation, check the coupling, and measure motor current against nameplate. A pump motor that keeps tripping its protection deserves its own diagnosis; see why motor overload relays trip.
• Hoses and valves: pinched or kinked hoses behind a machine, or a valve someone half closed, are embarrassingly common finds.
• Air locks: after fluid changes, bleed high points and let the pump purge with the reservoir cap arranged per the manual.
• Glycol concentration: too much glycol raises viscosity sharply, especially when cold, and the pump can no longer deliver design flow. Test with a refractometer rather than guessing.

Compressor short cycling and trips

Short cycling (compressor starting and stopping every minute or two) destroys compressors and usually signals one of three things: a refrigerant charge issue, a condenser problem driving high pressure cutouts, or an electrical fault. Watch one full cycle and note exactly what stops it: a pressure switch, the overload, or the controller.

• Low charge tends to trip low pressure cutouts; a blocked condenser or dead fan trips high pressure cutouts. Fault code meanings vary by model, series, and firmware, so confirm the code in the manufacturer manual rather than assuming.
• Electrical causes: pitted contactor contacts, a failing run capacitor on smaller units, or a tripping overload. Apply lockout/tagout and verify zero energy before opening any panel; chiller control circuits often stay live when the compressor is off.
• Burn hazard: compressor discharge lines run hot enough to burn skin on contact. Keep hands clear until the machine has cooled.

Never bridge a pressure switch or bypass an interlock to "keep production running." Those circuits are what stand between a nuisance trip and a destroyed compressor.

Freeze alarms and low temperature trips

A freeze alarm means the chiller believes the evaporator is about to ice up. Three causes cover nearly every case:

1. Setpoint below the safe limit for the fluid. Water-only chillers have a manufacturer minimum setpoint, commonly in the 7 to 10 degree C range. Running colder requires the right glycol mix, not a lower number in the controller.
2. Flow loss during operation. Low flow lets the fluid linger in the evaporator and overchill locally, even when the supply setpoint looks safe. Fix the flow fault first.
3. Bad temperature sensor. Verify against a calibrated reference thermometer at the supply port before replacing anything.

The freeze protection circuit is a safety, not a nuisance. Do not bypass it.

Water quality: scale, corrosion, and algae

Fluid problems cause slow, creeping failures: scale insulates heat exchange surfaces, corrosion particles clog strainers, and algae in open tanks turns the loop into soup. Three habits prevent most of it. Use an inhibited industrial glycol (not automotive antifreeze) at the concentration the manufacturer specifies, typically enough for freeze protection without wrecking viscosity. Keep makeup water sensible: soft or demineralized per the manual, never untreated hard water, and never repeated plain water top-ups that silently dilute the glycol. Cover open tanks and test the fluid on a schedule: concentration, inhibitor level, and pH.

Prevention rhythm that stops repeat faults

Almost every fault above is preventable with a boring, consistent routine: condenser coil cleaning, strainer inspection, fluid testing, a visual leak check, and a logged reading of supply temperature and flow. Slot these into a real preventive maintenance schedule at intervals matched to your environment; a dusty foundry needs coil cleaning far more often than a clean electronics shop. Logging the same two numbers every week is what makes drift visible early: a supply temperature creeping up over a month is a fouling condenser announcing itself before the alarm.

Measure it: make chiller failures visible

Treat every chiller fault as a downtime event with a cause code, logged against the machine it stopped, not as a war story in the break room. Track MTBF and MTTR for each chiller so a unit that fails every six weeks gets an engineering fix (better filtration, relocation out of the hot corner, a capacity review) instead of an endless cycle of resets. Trending those temperature and flow logs is also the on-ramp to condition-based maintenance, where the data decides when to intervene. And because a chiller fault surfaces as availability loss on the machine it serves, it belongs in your OEE numbers, where its true production cost becomes impossible to ignore.

Catch chiller-driven stops before they become chronic

Many chiller-related stops never make it into a logbook: the laser pauses for two minutes on a temperature warning, someone resets it, and the event evaporates. Fabrico is computer-vision-verified OEE plus closed-loop maintenance execution: cameras catch stops and micro-stops that manual logs and sensors miss, and maintenance work orders close the loop from detection to fix. If cooling faults keep nibbling at your uptime, book a Fabrico demo and see what your machines are actually doing.

Frequently asked questions

Why is my chiller running but not cooling?

The most common causes are a dirty condenser coil or air filter, a failed condenser fan, high machine room ambient, a fouled evaporator, or a refrigerant leak. Check heat rejection first: clean the coil, confirm fan airflow, then measure supply and return temperatures against the design delta before suspecting refrigerant.

What causes a chiller freeze alarm?

Usually a setpoint below the safe limit for the fluid, or a loss of flow during operation that lets fluid overchill in the evaporator. A failed temperature sensor is possible but less common, so verify the setpoint and flow before replacing parts, and never bypass the freeze protection circuit.

How often should condenser coils be cleaned?

It depends entirely on the air around the unit. In a dusty or oily environment, monthly inspection is reasonable; in a clean room, quarterly may be plenty. Log supply temperature weekly and let the trend tell you: a slow rise at constant load means the coil is fouling.

What glycol concentration should a process chiller run?

Whatever the manufacturer specifies for your lowest operating temperature, verified with a refractometer. Too little risks freezing the evaporator; too much raises viscosity, cuts heat transfer, and can trigger low flow alarms. Use inhibited industrial glycol, never automotive antifreeze.

Can I add refrigerant to a chiller myself?

No, unless you hold the required certification (F-Gas in the EU, EPA Section 608 in the US). A low charge also means there is a leak that must be found and repaired first, since simply topping up guarantees a repeat failure.