Two shops had already replaced the thermostat on this 2016 Subaru Outback. The owner still had a highway overheat problem. We pointed a thermal imaging camera at the radiator core and found the answer in under five minutes: the top tank was reading 192°F and the bottom third of the core was reading 108°F.
A functioning radiator does not have an 84-degree differential across its face. The radiator was 70% blocked by mineral scale and silicated coolant gel, it looked clean from the outside, and neither previous shop had tested it. Here is the full thermal scan, the flow rate data, and the physical autopsy of the core.
The Vehicle and What the Owner Was Told
Vehicle: 2016 Subaru Outback 2.5L (FB25 naturally aspirated), 89,400 miles at intake.
Reported symptom: Temperature gauge climbed to the red zone within 12 to 15 minutes of sustained highway driving at 70 mph. City driving, stop-and-go traffic, and idle temperatures were completely normal, sitting at the midpoint of the gauge without deviation.
Previous repair history: Two thermostat replacements at two separate shops over the preceding 14 months. The second shop also performed a chemical cooling system flush using a commercial oxalic acid product. Neither repair changed the highway behavior.
The highway-only symptom profile was the diagnostic fingerprint we needed before touching anything. A thermostat failure produces an overheating condition at idle and low speeds, not exclusively at highway speeds. Highway overheating with a normal idle temperature points to a cooling capacity problem: the system can maintain temperature under light load but cannot shed heat fast enough when the engine is working hard and demanding maximum coolant flow. The thermostat replacements were not wrong diagnoses exactly. They were diagnoses of the wrong system.
What the Thermal Camera Showed in Five Minutes
We pointed a Flir ONE Pro thermal imaging camera at the radiator face with the engine at full operating temperature after a 20-minute highway simulation on a chassis dyno at 70 mph equivalent load.
A functioning radiator transfers heat from the coolant to the air progressively across its core. The hottest coolant enters at the top tank, passes through the tubes, gives up heat to the fins, and exits cooler at the bottom tank. The temperature gradient across the face should be gradual and even: approximately 40 to 50 degrees from top to bottom across the full core height.
This is what we saw instead:
| Core Zone | Measured Surface Temp (F) | Expected Temp (F) | Delta |
|---|---|---|---|
| Top tank | 192 | 190 | Normal |
| Upper core (top 2 inches of tubes) | 184 | 178 | Normal |
| Mid-upper core | 171 | 165 | Slightly elevated |
| Mid core | 147 | 152 | Beginning to diverge |
| Lower mid core | 121 | 140 | Significant cold band |
| Lower core (bottom 3 inches) | 108 | 128 | Blocked zone |
| Bottom tank | 112 | 145 | 33 degrees below spec |
The thermal image showed a distinct horizontal cold band across the lower 40% of the radiator face. Hot coolant was entering the top tank, flowing through the upper tube rows, and essentially stopping there. The lower tube rows were receiving almost no flow. The core surface temperature in those zones matched ambient temperature more closely than operating coolant temperature.
This is what a 70% blocked radiator looks like on a thermal camera. At idle and city speeds, the partial upper-core capacity is enough to maintain temperature because thermal load is low. At highway speed under sustained engine load, the upper tubes cannot shed heat fast enough and the lower tubes are contributing almost nothing. The system hits a thermal ceiling and the gauge climbs.
Internal Flow Rate: Before and After
We installed a calibrated inline flow meter in the lower radiator hose and ran the engine at 2,000 RPM with the thermostat fully open. We measured flow through the clogged factory radiator, then measured again after installing a new OEM Denso unit for comparison.
| Condition | Measured Flow Rate (GPM) | OEM Specification Minimum |
|---|---|---|
| Clogged factory radiator | 4.2 GPM | 9.0 GPM |
| New OEM Denso radiator | 11.8 GPM | 9.0 GPM |
The clogged radiator was flowing 35% of the minimum required flow rate. The water pump was working correctly, generating normal pressure at the outlet. The restriction was entirely inside the radiator core itself.
The Autopsy: What We Found When We Cut It Open
We replaced the radiator before performing the physical inspection, so the cut-open unit was the removed original. We sectioned it with a rotary pipe cutter to expose the interior of the cross-flow tube bundles.
What a clean tube looks like: An open bore approximately 1.8mm wide running the full depth of the core, smooth inner walls, no restriction.
What we found in this radiator:
Six of the fourteen tube rows were completely blocked: a solid white-orange plug of calcium carbonate scale and silicated coolant gel running the full length of the tube. Four additional rows showed partial blockage ranging from 40% to 75% restriction. The remaining four rows were partially functional with minor scale deposits on the inner walls.
The scale had two distinct layers visible in cross-section. The inner layer was a hard white calcium carbonate crust approximately 0.3 to 0.5mm thick that had been deposited over multiple years. The outer layer filling the remaining tube bore was a soft orange-brown silicated gel: degraded coolant that had broken down its silicate inhibitor package, precipitated out of solution, and formed a sludge that hardened over time inside the tubes.
The calcium carbonate layer is the tap water signature. Mineral-laden tap water deposits calcium and magnesium salts on any heated surface they contact. Inside a radiator tube, those deposits accumulate with every heat cycle. A standard chemical flush with oxalic acid can dissolve light scale, but once calcium carbonate hardens and reaches the 0.3 to 0.5mm thickness we measured, a chemical flush cannot penetrate or dissolve it within a normal flush procedure. The two previous flushes had cleaned the soft surface deposits and left the hardened scale completely intact.
Did the Overheat Damage the Head Gasket?
The FB25 engine in the Subaru Outback has a documented sensitivity to coolant system failures. A 15-minute overheat event on this platform raises immediate head gasket concern, particularly given that Subaru’s EJ-series predecessors had widespread head gasket failures from far less thermal stress.
We performed a combustion leak test using a chemical block test kit (Lisle 75500) before any other work. The test draws air from the coolant reservoir through a detection fluid that changes color from blue to yellow in the presence of combustion gases. After three minutes of sampling with the engine at operating temperature:
Result: Negative. No color change detected.
The head gasket survived the overheat event. We attribute this to the fact that the overheat was a flow restriction failure rather than a sudden coolant loss failure. The engine did not run dry. Coolant was still circulating through the functional upper tube rows, providing partial cooling that prevented the head gasket from seeing the extreme sustained temperatures that cause deformation and seal failure.
We documented a compression test across all four cylinders as a secondary confirmation: 172, 174, 170, and 173 PSI, all within 4 PSI of each other and within specification. No combustion gas in the coolant and no compression variation: the short block and head gasket were intact.
Why the Flush Trap Caught Two Shops
The standard cooling system flush is not designed to remove hardened scale. It is designed to remove degraded coolant chemistry.
Commercial flush products (oxalic acid, citric acid, or proprietary blends) dissolve soft iron oxide deposits and flush out degraded silicate gel effectively. They do not dissolve calcium carbonate in its hardened, crystalline form within the contact time of a standard flush procedure, which is typically 20 to 40 minutes of idle circulation.
Industrial descaling requires stronger chemistry (hydrochloric or phosphoric acid solutions) and extended soak times of several hours, combined with mechanical agitation. Neither shop had the diagnostic data to know scale was the problem before performing the flush. Without a thermal scan or a flow rate measurement, a clogged radiator looks identical to a failing thermostat from a symptom-only perspective.
The Tap Water Root Cause
We found three notations in the vehicle’s service history where coolant level had been topped off, at two different quick-lube locations and one dealer visit. None of the three records specified the top-off fluid. At two locations, the repair order simply stated “added coolant to level.”
We contacted both quick-lube locations. One confirmed they use a 50/50 pre-mix, which would be acceptable. The other confirmed they top off with whatever is on hand, which at the time of the service in question was described as “distilled or tap water, depending on stock.” One tap water top-off in a system that is then heat-cycled repeatedly over 40,000 miles is enough to begin calcium carbonate scale formation in a radiator with narrow tube bores.
The Outback cooling system holds 7.4 quarts. A 1-quart tap water addition at a 60/40 coolant-to-water ratio reduces the distilled water fraction of the dilution and introduces minerals directly into the system. If the coolant pH is not tested and the inhibitor package is not refreshed on schedule (Subaru specifies Super Coolant replacement at 11 years or 137,500 miles for the first interval), the inhibitor breakdown accelerates mineral deposition.
How to Find a Blocked Radiator in Five Minutes With a $30 Tool
You do not need a thermal imaging camera to identify a blocked radiator. A basic infrared thermometer does the job when used correctly.
The procedure:
- Bring the engine to full operating temperature on a highway-load equivalent drive of at least 15 minutes.
- Immediately upon parking, open the hood and begin scanning the radiator face with the IR thermometer, holding the sensor 6 inches from the core surface.
- Scan in a grid pattern: three vertical columns (left, center, right) at five horizontal positions (top, upper-mid, center, lower-mid, bottom). Record all 15 readings.
- A uniform temperature gradient of 40 to 60 degrees from top to bottom indicates a functioning core.
- Any cold band, cold zone, or non-linear temperature drop greater than 25 degrees across adjacent measurement points indicates a blocked tube cluster in that zone.
The total time from parking to diagnosis: under five minutes. The total cost of the tool: $28 to $45 for a quality IR thermometer. This test would have identified the blockage on the first shop visit and saved the owner two thermostat replacements.
The Diagnostic Flow Table: Match the Symptom to the Fix
| Symptom | IR Temp Delta Across Core | Flow Rate Test | Likely Cause | Required Fix |
|---|---|---|---|---|
| Overheat at highway only, normal idle | Above 25 degrees differential | Below 7 GPM | Blocked radiator core | Radiator replacement |
| Overheat at idle and highway | Uniform, low overall temp | Normal flow | Thermostat stuck closed | Thermostat replacement |
| Slow warm-up, low idle temp | Uniform, low temp | Normal flow | Thermostat stuck open | Thermostat replacement |
| Overheating with coolant loss | Irrelevant | Irrelevant | Head gasket, hose, or water pump | Pressure test and leak trace |
| Overheating with no coolant loss, normal IR | Uniform gradient | Above 9 GPM | Water pump cavitation or ECT sensor | Flow and sensor test |
Cooling System Maintenance Rules to Prevent This
Water purity is non-negotiable. Only distilled water (resistivity above 1 megohm-cm) belongs in a cooling system. Tap water mineral content varies by region from 50 to 400 mg/L total dissolved solids. Any of that range is enough to deposit scale in narrow radiator tubes over 50,000 to 80,000 miles of heat cycling. Distilled water is $1.19 per gallon at any grocery store. It is not optional.
Coolant interval compliance matters for scale prevention, not just freeze protection. As coolant pH drops below 7.0 with inhibitor depletion, the acidic environment accelerates metal ion release from the aluminum components and the copper-brass radiator core. Those ions participate in scale formation. Subaru’s extended coolant interval is aggressive for high-mileage vehicles in hard water areas. We recommend pH testing at every coolant inspection, regardless of mileage interval.
Never top off with tap water, even once. One quart of tap water in a 7-quart system introduces enough mineral content to begin scale nucleation on heated tube surfaces within the first 5,000 miles of subsequent operation. If distilled water is unavailable in an emergency, use the pre-mixed 50/50 coolant from any parts store. It is pre-mixed with distilled water. The extra $4 over a jug of tap water is the cheapest preventive maintenance on this list.