If you want to confirm oxygen sensor (O2) health using live OBD2 data, you can do it reliably by combining two signals: the sensor’s live switching/response and the engine’s short-term and long-term fuel trims that react to that feedback. Next, you’ll learn a simple workflow that turns “random numbers on a scan tool” into a clear verdict: healthy, suspicious, or not the real problem.
Then, you’ll see what “normal” looks like in live data for upstream versus downstream sensors, including the upstream voltage switching pattern in closed loop and why the downstream sensor should usually look steadier. To illustrate, the article focuses on observations you can repeat at idle, at a steady cruise, and during quick rich/lean response checks.
Moreover, you’ll learn how to use fuel trims as your lie detector when O2 data looks confusing—like when a sensor appears “stuck” due to scan-tool refresh rate, or when a mixture problem makes a good sensor look bad. More specifically, you’ll use trim patterns across different engine loads to separate vacuum leaks, exhaust leaks, and fuel delivery issues from sensor failure.
Introduce a new idea: once you can interpret live O2 data confidently, you’ll also know when to stop testing and fix the underlying issue first—because some conditions make any O2 conclusion unreliable no matter how advanced your scanner is.
What does “healthy O2 sensor” mean when you’re looking at live OBD2 data?
A healthy O2 sensor is an exhaust feedback sensor that responds quickly and predictably to mixture changes so the ECU can keep the air-fuel ratio near target, which shows up as stable closed-loop control, normal switching/response patterns, and reasonable fuel trims.
Next, to avoid misdiagnosis, you need a clear definition of “health” that fits what live data can actually prove.
A “healthy” verdict from live data is not “the number looks nice.” It is the combination of three behaviors that reinforce each other:
- Correct feedback behavior: the sensor output changes when the mixture changes.
- Correct control behavior: the ECU reacts to that feedback and adjusts fueling (you see it in STFT).
- Correct overall result: the engine settles into stable control with trims that are not compensating excessively.
This definition matters because O2 sensors do not exist in isolation. A sensor can be “telling the truth” about a problem you do not want to hear—like an exhaust leak pulling in oxygen ahead of the sensor, or a vacuum leak creating a real lean condition. In those cases, the sensor looks “lean,” but it is healthy.
Which live-data PIDs should you monitor to confirm O2 sensor health?
There are 7 core PIDs to monitor to confirm O2 sensor health: O2/AF sensor signal, STFT, LTFT, loop status, coolant temperature, engine speed/load, and airflow/pressure—because those are the minimum set that lets you connect cause (mixture change) to effect (sensor response) to reaction (trim correction).
Then, once you watch them together, the sensor data becomes meaningful instead of isolated.
Use this checklist as your baseline data stream:
- O2 sensor or A/F sensor PID
- Narrowband O2 is typically displayed as voltage.
- Wideband often displays lambda/equivalence ratio or current (tool-dependent).
- Short-Term Fuel Trim (STFT)
- The ECU’s immediate correction based on feedback.
- Long-Term Fuel Trim (LTFT)
- The ECU’s learned correction over time.
- Fuel system status / loop status
- You need closed loop for O2 feedback evaluation.
- Coolant temperature (ECT)
- Confirms the engine is warm enough to be in normal control.
- RPM + calculated load (or MAP/MAF)
- Lets you compare idle vs cruise vs light acceleration apples-to-apples.
- Throttle position (optional but helpful)
- Helps interpret transient events like snap throttle or decel fuel cut.
When your scanner allows it, graph O2 and STFT together. Graphs are not just prettier—they show timing and response speed, which is often the whole diagnosis.
What’s the difference between upstream (pre-cat) and downstream (post-cat) readings in live data?
Upstream wins for fuel control feedback, downstream is best for catalyst monitoring, and comparing them is optimal for catching “false sensor blame” because they should behave differently in a functioning system.
However, you need to understand that difference before you declare a sensor bad.
Use this mental model for Upstream vs downstream O2 sensor differences:
- Upstream (B1S1 / B2S1, pre-cat)
- Primary feedback sensor for mixture control.
- In closed loop, it should show active switching or active correction behavior.
- Downstream (B1S2 / B2S2, post-cat)
- Monitors catalytic converter efficiency by observing how much the converter smooths oxygen fluctuations.
- It should usually look steadier than the upstream signal during steady driving.
If you treat a downstream sensor like an upstream sensor, you will “fail” good parts. If you treat an upstream sensor like a downstream sensor, you will miss control problems.
Is your scan tool setup good enough to trust the O2 live-data conclusions?
Yes—your scan tool setup is good enough to trust O2 live-data conclusions if it confirms closed loop, shows the correct sensor PID, has adequate refresh/graphing, and the engine is fully warmed—because those four conditions prevent the most common “scanner-created” misreads.
Next, you’ll lock down the prerequisites so your conclusions come from the engine, not from tool limitations.
Here are the three biggest reasons setup matters:
- Closed-loop control is the context: without it, the ECU may ignore the sensor, and “flat” readings are meaningless.
- Refresh rate controls what you can see: a slow scanner can make switching look lazy or stuck.
- Wrong PID selection creates fake symptoms: selecting the wrong bank/sensor can make you chase a fault on the wrong side of the engine.
What conditions must be true before you evaluate O2 live data?
There are 6 conditions that must be true before you evaluate O2 live data: engine warmed, closed loop active, stable idle available, no active misfire, no obvious exhaust leaks, and the correct bank/sensor selected—because each one directly affects oxygen content or how the ECU uses feedback.
Then, once these conditions are met, your observations become repeatable and diagnostic.
Use this pre-check list:
- Coolant temperature at normal operating range (not “barely warm”)
- Fuel system status shows closed loop
- No active misfire behavior (misfire creates oxygen in exhaust and lies to the sensor)
- No obvious exhaust leak noise near manifolds/pipe joints (leaks pull in oxygen)
- No major vacuum leak symptoms (high idle, severe trims, hissing)
- Correct PID: confirm you’re watching B1S1 vs B1S2, etc.
This is also where many people see Bad O2 sensor symptoms that are not actually sensor failure: rough idle, poor fuel economy, hesitation, and a check engine light can also come from vacuum leaks, MAF errors, fuel pressure issues, or misfires.
How can scan-tool refresh rate and smoothing make a healthy sensor look bad (or vice versa)?
Scan-tool refresh rate and smoothing can make a healthy sensor look bad by averaging fast switching into a slower, flatter line, and they can make a bad sensor look okay by hiding dropouts or delays—because the tool changes the shape of the signal you are judging.
Specifically, you should judge response speed only when the data rate is high enough to show it.
Common “tool artifacts” include:
- Stair-step graphs: the tool is sampling slowly, so the waveform looks chunky.
- Flat-ish midline at ~0.4–0.6V: the tool may be averaging; you might still be in closed loop.
- Lag between throttle change and displayed response: the tool is late, not the sensor.
Practical fix: if your scanner can’t graph quickly, log fewer PIDs at once. The more PIDs you display, the slower many budget tools update.
What “normal” voltage switching looks like for a narrowband upstream O2 sensor?
Normal narrowband upstream O2 switching is rapid cycling between lean and rich indications during closed loop, reflecting the ECU’s constant micro-adjustments around stoichiometry, and it typically appears as repeated voltage swings rather than a steady line.
Next, you’ll learn what “normal” looks like in patterns, not just numbers.
A narrowband upstream O2 sensor is designed to report whether the mixture is richer or leaner than target—not the exact air-fuel ratio. So in closed loop, you should expect the ECU to “hunt” slightly rich then slightly lean, and the sensor reports that cycling. That is not instability; it is control.
A healthy upstream pattern usually has these traits:
- Active oscillation once fully warm and in closed loop
- Consistent rhythm at steady conditions (idle or cruise)
- Fast reaction to intentional changes (snap throttle, brief enrichment/leaning)
If the signal is truly stuck high or low, you still need to verify whether that is a sensor fault or a real mixture condition.
Should a healthy upstream O2 sensor switch at idle, and what patterns count as normal?
Yes—a healthy upstream O2 sensor should switch at idle in closed loop, and normal patterns include repeated lean-to-rich cycling, quick recovery after small throttle changes, and stable fuel trims that hover near zero correction at steady idle.
However, because idle is sensitive to small leaks, you must confirm idle stability before you judge the waveform.
At idle, a common healthy pattern is “busy” switching that looks like:
- The voltage repeatedly moves between low and high zones (tool-dependent scaling).
- STFT dances around small positive/negative values as the ECU corrects.
- The pattern repeats with a recognizable cadence when the engine is stable.
If idle is unstable (misfire, vacuum leak, IAC/ETC fluctuations), the sensor will look chaotic because the exhaust oxygen content is chaotic.
What live-data patterns indicate a “lazy” O2 sensor versus a real mixture problem?
A lazy O2 sensor wins for “slow response and low switching activity,” a real mixture problem is best identified by “fuel trims and corroborating PIDs,” and comparing both is optimal because trims tell you whether the ECU agrees with the sensor’s story.
Meanwhile, this is where many people replace sensors that were correctly reporting a different fault.
Use this comparison logic:
Lazy sensor pattern (sensor problem)
- Switching slows down even when conditions are steady and warm.
- The sensor reacts late during a deliberate rich/lean event.
- STFT may become less responsive or oscillate oddly because feedback is delayed.
Real mixture problem pattern (engine problem)
- O2 looks lean and STFT stays significantly positive (ECU adding fuel).
- LTFT trends positive over time.
- The condition changes with load (idle vs cruise), which often points to leaks or fuel delivery.
If you see persistent lean indications plus high positive trims, suspect causes before you jump to oxygen sensor replacement, such as intake leaks, unmetered air, or exhaust leaks upstream of the sensor.
How do you confirm O2 sensor health with quick rich/lean response tests using live data?
Confirming O2 sensor health with quick rich/lean tests means performing 3 repeatable steps—stabilize in closed loop, induce a brief rich or lean change safely, and verify the sensor and fuel trims respond quickly—so you can prove the sensor reacts to reality.
Next, you’ll learn safe test ideas that work with live data without turning diagnostics into damage.
A response test is powerful because it turns “interpretation” into “verification.” You are no longer guessing what the waveform means—you are asking the engine a question and watching whether the sensor answers correctly.
Safety note baked into the method: the goal is a brief change, not a long rich run that overheats a catalyst or a long lean condition that risks misfire. Keep tests short and controlled.
What is a safe, practical rich test and what should you see in O2 voltage and STFT?
A safe, practical rich test is a brief enrichment event—often a quick snap throttle from idle or a short controlled enrichment method—where the upstream O2 voltage should move rich quickly and STFT should respond by pulling fuel back as the ECU restores target mixture.
Then, once you see the paired response, you have direct evidence the sensor is alive and the ECU is listening.
A simple rich-style check you can do without specialty equipment:
- Warm engine fully and confirm closed loop.
- Stabilize at idle for 20–30 seconds so the baseline is steady.
- Snap the throttle briefly (a quick, moderate blip) and watch:
- O2 should swing rich shortly after the event.
- STFT should change rapidly as control resumes.
Interpretation tips:
- If the sensor responds quickly but the ECU struggles to recover, the issue may be fueling control, not the sensor.
- If the ECU reacts (STFT changes) but the sensor does not, suspect sensor signal issues or PID mismatch.
What is a safe, practical lean test and what should you see in O2 voltage and STFT?
A safe, practical lean test is a brief event that increases oxygen in the exhaust without prolonged stress—often a controlled light-load condition change—where the upstream O2 should indicate lean quickly and STFT should move positive as the ECU adds fuel to return to target.
However, because “lean” can be created by leaks and misfires too, you should keep the test controlled and confirm stability first.
A practical lean-style check on many vehicles is observing decel fuel cut and recovery during a safe road test:
- Cruise steadily at a safe speed.
- Lift off the throttle smoothly (no braking panic) and watch for decel behavior.
- As the vehicle transitions back to light throttle, observe:
- O2 changes as combustion returns.
- STFT reacts quickly as closed loop re-engages.
The key is not the exact voltage number—it’s the timing: the sensor should respond promptly when operating conditions change.
Can fuel trims confirm O2 sensor health even if O2 voltage looks “weird”?
Yes—fuel trims can confirm O2 sensor health even if O2 voltage looks weird because STFT is the ECU’s direct reaction to feedback, LTFT shows learned correction over time, and consistent trim behavior across loads can validate that the sensor is providing usable information.
Moreover, trims can also prove the opposite: that the ECU is compensating for a real problem the sensor is correctly reporting.
Here’s how trims “verify” the sensor:
- If STFT moves rapidly in response to small mixture changes, the feedback loop is functioning.
- If LTFT is reasonable and does not drift aggressively, the system is not constantly fighting a false signal.
- If trims are extreme but consistent with other symptoms (vacuum leak at idle, weak fuel delivery under load), the sensor may be telling the truth.
This is also where you connect diagnostics to “what people actually search for,” like Common O2 sensor codes (P0130–P0167 families, heater codes, slow response codes, and catalyst efficiency codes). The code points you to which bank/sensor the ECU mistrusts, but live data tells you why.
How do fuel trims validate (or contradict) O2 sensor readings in real driving conditions?
Fuel trims validate or contradict O2 readings by showing how much correction the ECU must apply to satisfy feedback across different loads, so a healthy sensor usually pairs with modest, logical trim behavior while false lean/rich conditions create trim patterns that reveal the real fault.
Next, you’ll use trim patterns like a map: they do not just say “lean” or “rich,” they say where and when.
Think of trims as the ECU’s diary:
- STFT is what the ECU is doing right now.
- LTFT is what it has learned it must do to keep things normal.
A fast diagnostic trick is to compare trims at:
- Idle
- ~2500 RPM no-load (if safe and appropriate)
- Steady cruise
- Light acceleration
When trims change dramatically by condition, the cause is often not the O2 sensor—it’s airflow, vacuum, exhaust integrity, or fuel delivery changing with load.
Which fuel-trim patterns point to vacuum leaks, fuel pressure issues, or exhaust leaks (not a bad sensor)?
There are 4 main fuel-trim patterns that point away from a bad O2 sensor and toward vacuum leaks, fuel pressure issues, exhaust leaks, or metering errors—because each fault changes oxygen content differently depending on engine load.
Then, once you match the pattern, you stop guessing and start testing the right system.
Use these pattern groups:
- Lean mostly at idle (STFT/LTFT positive at idle, improves at higher RPM/load)
- Often points to vacuum leaks or unmetered air.
- Idle is high-vacuum, so leaks matter most there.
- Lean under load / at cruise (positive trims worsen with acceleration or sustained speed)
- Often points to fuel delivery issues (weak pump, restricted filter, low pressure) or under-reporting airflow.
- Load demands fuel; shortages show up here.
- Intermittent lean with noisy O2 activity and strange trim spikes
- Can point to exhaust leaks ahead of the sensor or wiring issues.
- Fresh air entering the exhaust can create false lean signals.
- Rich corrections (negative trims) with soot smell and poor economy
- Can point to leaking injectors, high fuel pressure, evap purge faults, or misfire-related oxygen confusion.
- In these cases, the O2 sensor may still be working perfectly.
If your trims tell a consistent story and the O2 sensor responds quickly to changes, do not replace the sensor just because the waveform looks ugly. That is how unnecessary oxygen sensor replacement happens.
How do you compare Bank 1 vs Bank 2 data to spot sensor-specific problems?
Bank comparison wins for isolating “one side lying,” single-sensor comparison is best for verifying response behavior, and using both is optimal because bank-to-bank symmetry reveals whether the problem is shared (engine-wide) or isolated (sensor-side).
However, you must compare banks under the same operating condition to make the conclusion valid.
Do this bank comparison correctly:
- Hold the engine at the same condition (idle stabilized, or steady cruise).
- Compare B1S1 vs B2S1 (upstream sensors) and their related trims if your tool supports bank-specific trims.
- Look for asymmetry:
- One upstream sensor is slow or stuck while the other behaves normally.
- One bank shows consistently different trim behavior.
Asymmetry often points to:
- One sensor/wiring issue
- One-side exhaust leak
- One bank running differently due to injector or intake leak on that side
Symmetry often points to:
- Shared fueling issue
- Shared MAF/airflow measurement error
- Shared fuel pressure issue
Is it the O2 sensor or the catalytic converter—how can you tell from upstream vs downstream live data?
Upstream data wins for diagnosing mixture control, downstream behavior is best for evaluating catalyst smoothing, and comparing upstream vs downstream is optimal for telling whether the catalytic converter is doing its job or whether you’re chasing the wrong component.
Next, you’ll apply that comparison so you don’t confuse a catalyst problem with a sensor problem.
In a typical system:
- The upstream sensor swings because the ECU is controlling mixture.
- The catalytic converter stores and releases oxygen to smooth those swings.
- The downstream sensor “sees” the smoothed result.
If the converter is healthy, the downstream signal is usually less active in steady-state conditions. If the converter is degraded, the downstream may start to resemble the upstream more closely.
Should the downstream O2 sensor switch like the upstream sensor?
No—the downstream O2 sensor should not switch like the upstream sensor during steady operation because a functioning catalytic converter reduces oxygen fluctuation, which makes the post-cat signal steadier, and the downstream sensor’s primary job is monitoring converter efficiency.
However, brief changes can still cause movement, so you judge patterns under steady conditions, not during transients.
If your downstream looks extremely “busy” at steady cruise and closely mirrors upstream, treat that as a catalyst-efficiency clue—not automatically a downstream sensor failure.
What patterns suggest catalyst inefficiency versus a faulty downstream sensor?
Catalyst inefficiency wins for “downstream mirrors upstream under steady conditions,” a faulty downstream sensor is best indicated by “erratic or non-responsive signal that doesn’t match operating changes,” and comparing both against trims is optimal because trims reveal whether control is stable while monitoring looks abnormal.
Meanwhile, this is where live data prevents expensive mistakes.
Use this pattern comparison:
Likely catalyst inefficiency
- Upstream shows normal control switching.
- Downstream shows similar switching amplitude and rhythm during steady cruise.
- Trims are not extreme, suggesting mixture control is not the main issue.
Likely downstream sensor or circuit issue
- Downstream signal is flat, drops out, spikes randomly, or behaves inconsistently with temperature/load.
- The behavior may appear disconnected from what upstream and trims are doing.
- You may see heater-related codes or signal circuit codes in addition to efficiency codes.
Evidence matters here because research on exhaust aftertreatment shows how much can be inferred from upstream/downstream lambda responses. According to a study by Clemson University from the International Center for Automotive Research (CU-ICAR), in 2016, researchers reported that the post-catalyst (downstream) lambda sensor response can be used as an indicator of three-way catalyst aging, supporting the idea that downstream response behavior carries meaningful catalyst information in diagnostics.
When should you stop live-data testing and fix other problems first?
Yes—you should stop live-data testing and fix other problems first if misfires, exhaust leaks, unstable temperature control, or sensor power/wiring faults are present because those conditions distort exhaust oxygen or invalidate closed-loop feedback, making O2 data conclusions unreliable.
Next, you’ll learn the “stop signs” that protect you from diagnosing noise.
Here are three reasons to stop and repair first:
- Misfires and incomplete combustion inject oxygen into the exhaust, creating false lean indications.
- Exhaust leaks ahead of the sensor pull in fresh air, making the sensor “report lean” even when fueling is correct.
- Open-loop operation prevents meaningful feedback evaluation, because the ECU may not use the sensor at all.
Which symptoms mean live O2 readings are unreliable until repaired?
There are 7 symptoms that make live O2 readings unreliable until repaired: active misfire behavior, loud exhaust leaks, coolant temperature not reaching normal, unstable idle, heater circuit faults, obvious wiring damage, and severe trim extremes that don’t stabilize—because each one breaks the normal relationship between exhaust oxygen, sensor output, and ECU correction.
Then, once these are corrected, your live-data workflow becomes trustworthy again.
Look for these “stop signs”:
- Misfire symptoms: shaking at idle, misfire counters rising, misfire codes
- Exhaust leak signs: ticking near manifolds, soot marks, strong odor near joints
- Thermostat/temperature issues: never reaches normal temp, frequent open-loop behavior
- Heater circuit issues: heater codes, unusually slow sensor warm-up, persistent open loop
- Wiring/connectors: melted harness near exhaust, intermittent PID dropouts
- Extreme trims: STFT pinned high/low with LTFT marching in the same direction
- Recent changes: after repairs, unplugged vacuum lines or intake boots
This is also the right time to interpret “what the car is telling you” through symptoms rather than just numbers. Classic Bad O2 sensor symptoms can include reduced fuel economy, rough running, hesitation, and failed emissions readiness, but those symptoms overlap with many faults—so treat them as a reason to test, not proof of sensor failure.
What changes when your vehicle uses a wideband A/F sensor instead of a narrowband O2 sensor?
Wideband wins for precise mixture measurement across a broader range, narrowband is best for simple rich/lean switching near stoichiometry, and understanding which one you have is optimal because wideband live data often uses lambda/equivalence ratio instead of “voltage switching,” changing how you confirm health.
Next, you’ll expand from the main workflow into sensor-type-specific interpretation so you don’t apply the wrong rulebook.
A wideband (A/F or UEGO) sensor can show:
- Lambda (target near 1.0 at stoichiometric)
- Equivalence ratio
- Sensor current or a calculated AFR value (tool-dependent)
A narrowband (HEGO) sensor usually shows:
- Switching voltage behavior used for stoichiometric control
This difference matters because many DIYers see a wideband PID and assume it should behave like a narrowband voltage trace. That mismatch creates false diagnoses and unnecessary parts swapping.
How do you confirm wideband sensor health using lambda, equivalence ratio, or A/F current?
Confirming wideband sensor health means verifying that lambda (or equivalence ratio) tracks target quickly, stabilizes under steady conditions, and responds promptly to controlled changes while trims remain logical—because wideband health shows up as accurate tracking and responsive control rather than classic 0.1–0.9V switching.
Then, once you confirm tracking behavior, you can trust the wideband signal for deeper fueling diagnostics.
Use these checkpoints:
- Warm engine + closed loop still apply.
- At steady cruise, lambda should remain near target with small corrections.
- During light throttle changes, lambda should move appropriately and recover quickly.
- Trims should not be extreme unless another fault is truly present.
If the wideband signal is noisy, stuck, or slow while trims behave strangely, suspect sensor contamination, wiring, or heater issues before you accuse the ECU.
What is Mode $06 and can it confirm O2/A/F sensor performance without guessing from graphs?
Mode $06 is an onboard diagnostic test-reporting mode that can expose monitored test results for emissions-related components, and it can sometimes confirm O2/A/F sensor performance by showing pass/fail-style test values without relying only on waveform interpretation.
However, Mode $06 support varies widely by make/model and by scan tool, so it is a confirmatory tool rather than a universal solution.
When Mode $06 is available for oxygen sensor monitoring, it can help you:
- Confirm whether the ECU considers the sensor responding within expected limits
- Compare measured values to thresholds
- Reduce guesswork when the waveform is hard to interpret
If your scanner supports it, Mode $06 is especially useful when you suspect “borderline” performance and want an ECU-level view of test outcomes.
Which rare issues can mimic a bad O2 sensor in live data ?
There are 4 rare issues that mimic a bad O2 sensor in live data: exhaust leaks ahead of the sensor, sensor contamination, wiring resistance/ground problems, and PID update lag—because each one changes the displayed signal without reflecting true mixture or true sensor element behavior.
Then, once you rule these out, your live-data conclusion becomes much more confident.
- Exhaust leak ahead of sensor
- Mimic: persistent lean indication and high positive trims.
- Rule-out: inspect for soot marks, listen for ticks, smoke test exhaust if needed.
- Sensor contamination (coolant, silicone, fuel additives in excess)
- Mimic: slow response, biased readings.
- Rule-out: address root cause (coolant burning, sealant misuse) before replacing again.
- Wiring/ground resistance
- Mimic: dropouts, stuck readings, heater-related weirdness.
- Rule-out: inspect harness routing near hot exhaust, verify connector integrity.
- PID update lag / CAN latency
- Mimic: “lazy sensor” that is actually a slow tool.
- Rule-out: graph fewer PIDs, compare with another tool, or log at higher rate.
Should you replace the O2 sensor if live data looks borderline but fuel trims are normal?
No—you should not replace the O2 sensor just because live data looks borderline if fuel trims are normal, upstream/downstream behavior matches expectations, and response tests show timely reaction, because normal trims indicate the ECU is achieving target control and the sensor feedback is usable.
In addition, when replacement is justified, you should tie it to evidence like slow response codes, heater faults, dropouts, or repeatable response delays rather than aesthetics of a waveform.
This is the practical replacement rule:
- Replace when the sensor fails a response test, sets relevant codes repeatedly, or shows clear circuit/heater issues.
- Do not replace when the system is controlling normally and only the display “looks odd.”
When replacement is necessary, plan the job as a complete repair decision, not just parts swapping: confirm the correct bank/sensor, address any exhaust leaks, and treat the repair as an emissions system fix. That is how you prevent repeat failures after oxygen sensor replacement—especially when the original issue was a leak or contamination upstream.
Evidence (if any)
According to a study by Clemson University from the International Center for Automotive Research (CU-ICAR), in 2016, researchers reported that the post-catalyst (downstream) lambda sensor response can be used as an indicator of three-way catalyst aging, supporting the diagnostic value of upstream/downstream response comparison when evaluating sensor and catalyst behavior.
