Your Thermostat, Water Pump, or Alternator Probably Wasn't the Problem—Your Diagnosis Was

The return bin at our inspection station is the most honest place in the plant.

Every week—or almost every week, some weeks are worse—we get a fresh batch of parts that customers brought back. Thermostats. Water pumps. Ignition coils. Alternators. Each one has a story. "Doesn't work." "Engine still overheats." "Check engine light came right back."

Then we test them. I've been doing this since 2019, reviewing around 200 unique returned units per year as a quality and brand compliance manager. And here's what I find over and over: most of these parts are fine.

Not "fine for the price." I mean factory-spec, within-tolerance, nothing-wrong-with-this-part fine. Last month, 73% of the returned units we tested passed inspection. Some of the rest were damaged during installation—a cracked housing, a reused gasket, an overtightened bolt.

The parts weren't bad. The diagnosis was.

"Most of the 'failed' parts on my bench aren't failed. They were diagnosed wrong."

The usual suspects (and why they're usually innocent)

Let me walk through the scenario I see most often, because it involves several terms people type into Google: Gates 33945 O.E. type engine coolant thermostat, Gates water pump, clogged catalytic converter, and 2017 Chevy Equinox alternator.

Say you drive a 2017 Equinox. The temperature gauge starts creeping up, or the check engine light comes on. You do what any reasonable person does: search the symptoms, find a part that fits, and buy it.

For overheating, the Gates 33945 O.E. type engine coolant thermostat is a natural first suspect—it's a legitimate OEM-spec replacement for several GM engines, and a stuck-closed thermostat will absolutely cook an engine. So you swap it in, bleed the cooling system, and go for a drive. The needle still climbs.

Next, circulation. A Gates water pump with a worn impeller has caused exactly the same symptom: coolant that isn't moving. You install it, fresh coolant and a new belt while you're at it. The engine still runs warmer than it should.

At that point, the mind goes to head gaskets and a $2,000 repair. But there's another possibility most drivers don't think about until it's too late: a clogged catalytic converter. Restricted exhaust creates backpressure that smothers the engine's ability to breathe, and one of the symptoms is exactly that—temperature creep, especially under load or on the highway. The cooling system was fine. The parts you installed were fine. You were solving a theory, not the problem.

The alternator that wasn't (and the coil that was)

Or take the 2017 Chevy Equinox alternator scenario—a classic. Battery light on, dim dash lights, maybe a stall at a stoplight. Textbook "alternator is dying" behavior.

Except on many GM vehicles from that generation, the charging system is managed by the engine control module using a variable-voltage strategy. The alternator only produces what the ECM commands. If the battery is fading, or a ground strap is corroded, or the battery current sensor is reading wrong, the whole system can mimic a dead alternator beautifully. People install a new alternator, the light comes back three days later, and the old alternator gets returned to us. It tests fine—I've personally run output tests on over a hundred GM alternators that came back this way, and at least two-thirds were completely within spec.

The numbers said the alternator was fine—14.1 volts at the battery, full output under load. My gut said the customer wasn't going to believe us. Turns out the real issue was a failing battery that killed the charge cycle overnight. A load test on the battery would have caught it in 10 minutes, for free at most parts stores.

Then there's the ignition coil I see repeatedly in the same story: a car with a misfire dumps raw fuel into the exhaust. That unburned fuel raises converter temperatures and eventually destroys the catalyst. If you replace the converter without fixing the coil, the new converter will die the same way. The coil you didn't replace was the root cause; the converter was collateral damage. I've seen this cycle a dozen times this year alone.

Why this keeps happening

I don't think drivers are careless. I think three things set them up for this pattern:

One: symptoms overlap more than people expected. Overheating can come from the thermostat, the water pump, a clogged converter, a radiator cap that lost its seal, a fan clutch, air in the system, or a head gasket. The check engine light could be ignition, fuel, emissions, or a loose gas cap. A battery light could be the alternator, the battery, a ground, or a strategy that's working exactly as designed. Picking the most memorable cause and replacing it is a guess with a credit card attached.

Two: the real problem can damage new parts. This is the one that frustrates me as an inspector. A new converter installed downstream of a misfiring engine isn't a repair; it's a countdown. The new part fails because the underlying condition was never addressed, and then the new part gets blamed.

Three: testing doesn't feel like real work. In our plant, every shipment is checked against a written spec list before it leaves. No checklist, no shipment. That discipline isn't bureaucracy—it's that catching a defect in pre-shipment review costs 1/10th of what it costs us if it reaches the customer. The same math applies in your driveway. A 10-minute test feels slow; a weekend of rework feels slower.

What a misdiagnosis actually costs

Let me run the numbers on that 2017 Chevy Equinox, using approximate U.S. aftermarket pricing as of March 2025 (verify current figures—prices shift):

• Thermostat replacement (Gates 33945 or OE-equivalent, installed): $150–220
• Water pump replacement (Gates water pump, installed): $230–380
• Alternator replacement: $320–480
• Catalytic converter replacement: $550–1,400
• Ignition coil and plugs: $180–300

Guess your way through all five and you're out $1,430 to $2,780—and you still might not have fixed the car if the root cause was wiring or a battery.

Last year we received a "defective" catalytic converter from a customer who had also replaced the thermostat, the water pump, and two oxygen sensors before calling us. 4,000 miles on the converter, melted substrate. Root cause: a single ignition coil. He spent roughly $2,800 across all those repairs. The coil alone would have been about $45 and an hour of his time. The upside of him calling us was that we could tell him the converter was killed, not defective. The downside was that his first interaction with our brand was a warranty claim on a part he never needed in the first place. (Fortunately, we covered it.)

Ten minutes of testing beats five days of rework

What should you actually do? Test before you replace. Not after. Before. Here's the short list—most of it only needs a $20–40 multimeter and a little patience.

How to test an ignition coil with a multimeter: Disconnect the coil and set the multimeter to ohms. Check primary resistance across the two small connector terminals—typically 0.3 to 1.0 ohms for coil-on-plug units (verify against the vehicle's service spec). Then check secondary resistance from the high-tension terminal (the spring-loaded contact that touches the spark plug) to the positive pin—usually 6,000 to 15,000 ohms. Out of range means replace. In range means the coil is probably not your problem. (Rare heat-sensitive failures can pass a static test; if you have access to an oscilloscope, that's the definitive check.)

Check for a clogged catalytic converter with a vacuum gauge: Connect the gauge to the intake manifold, run the engine at idle—the vacuum should be steady, typically 15–20 inHg. Raise it to 2,500 rpm and hold. A severely restricted converter causes vacuum to drop slowly at steady throttle. An exhaust backpressure gauge threaded into the upstream O2 sensor port is even better: below 1.25 psi at idle is normal; above 2 psi, the converter is clogged. (Don't mix up the O2 port with anything else—ask me how I know.)

Test the alternator at the battery, not at the alternator: With the engine running and headlights on, a healthy system should show 13.8–14.5 volts at the battery terminals. If it does, the alternator is doing its job—so check the battery condition and the ground strap before considering replacement. Most parts stores will load-test the battery for free, which eliminates the other half of that equation.

Test the new thermostat before installing it: Put it in a pot of hot water with a kitchen thermometer. The valve should start opening at the temperature stamped on the flange, and open fully within about 20°F above that. If it opens early, late, or not at all—return it before installation, not after. It's cheap insurance, and it takes one partially-read recipe's worth of time.

This approach works well for us in inspection because we have a standardized bench setup and years of failure data. If you're working in a gravel driveway with a cheap multimeter, the process is the same—just less clean. And in my experience, which is admittedly focused on the parts that come back to me rather than the ones that stay in cars, verified diagnosis rarely sends a customer down the wrong path.

Testing first isn't glamorous. It doesn't feel as satisfying as swapping a part and crossing your fingers. But the part doesn't care what you believe about it. It only cares whether it's within spec. That's true for a $45 ignition coil, a $250 alternator, and every component that lands on my bench.