Diagnostics & Drivability (OBD-II)

Knowing what each mode of data offers — and what it can't tell you — is the difference between a tech who reads a code and replaces a part, and a diagnostician who lets the vehicle tell its own story.

The modes that matter: freeze frame and Mode $06

When the ECU sets a trouble code, it doesn't just store the code — Mode $02, the freeze frame, captures a snapshot of engine parameters at the exact moment the fault occurred: RPM, engine load, coolant temperature, fuel trims, vehicle speed, and more. This is your reproduction recipe. If the freeze frame shows the code set at 2,400 RPM, 65 percent load, and full operating temperature, you now know exactly where to drive the vehicle to make the fault happen again. The car is telling you where it was operating when it failed — use that instead of guessing.

Mode $06 goes a level deeper. It exposes the raw test results the OBD-II monitors actually use to decide pass or fail, along with the pass/fail limits: catalyst efficiency ratios, misfire counts per 200 revolutions, EVAP leak test pressures. This is invaluable for marginal or intermittent failures. Take an intermittent P0420 catalyst code: Mode $06 shows the actual efficiency ratio the ECU calculated during the catalyst monitor's test. If the measured ratio sits barely above the failure threshold — say 0.05 against a 0.02 threshold — the converter is degrading but not fully dead, and you can advise the customer on repair timing. A ratio deep in the fail range, like 0.001, means the cat is completely gone. That detail changes the whole repair conversation.

Generic vs. enhanced data

Generic OBD-II is the legally mandated subset of diagnostic data, and it is almost entirely emissions-related — the PIDs and codes every scan tool can read on every vehicle. Manufacturer-enhanced data is everything else: ABS, airbag, transmission, body electrical, and thousands of manufacturer-specific PIDs accessed through manufacturer protocols or capable aftermarket scan tools. If you're diagnosing anything beyond emissions on a modern vehicle — a shift complaint, a body electrical gremlin, a chassis fault — you need enhanced scan tool access. A generic code reader simply cannot see those systems.

Pending codes, confirmed codes, and monitor readiness

OBD-II requires most faults to occur on two consecutive drive cycles before the check engine light illuminates. A fault detected on the first drive cycle is stored as a pending code — no CEL yet. If it repeats on the next drive cycle, it becomes a confirmed code and the light comes on. Pending codes are early warnings, and they're gold for catching intermittent problems before they mature into a full complaint. Always check for them.

Monitor readiness is a separate concept that trips up a lot of techs. The OBD-II system runs a series of self-tests — monitors for the O2 sensors, catalyst, EVAP system, and others — automatically, but only when specific drive conditions are met. Readiness status tells you whether each monitor has completed since the last code clear or battery disconnect. This matters because clearing codes resets all monitors to not-ready, and most states require monitors to show ready before the vehicle can pass an emissions inspection. So after you clear codes following a repair, the vehicle must complete a drive cycle — a specific sequence of speeds, temperatures, and cruise times published in the manufacturer's service data — before it can be emissions tested. A vehicle that fails inspection for not-ready monitors hasn't failed anything; the tests just haven't run yet.

Applying it: chasing an EVAP small leak

A P0442 (EVAP small leak) is a good example of letting the system guide you. Start with the gas cap — verify it's tight and the gasket seals, since that's the cheapest possible fix. But don't stop there and don't blindly replace it. Small EVAP leaks are usually impossible to locate by eye or ear, so the professional tool is a smoke machine: pressurize the EVAP system with smoke and look for where it escapes. Common leak points are the purge and vent valves, the charcoal canister, hoses, and fuel tank connections. Check TSBs too — many vehicles have documented, model-specific leak locations that can save you an hour of hunting.

Read them correctly and the engine tells you whether it has a vacuum leak, a lazy MAF, a leaking injector, or a fuel pressure problem — often before you ever pick up a wrench.

What STFT and LTFT actually mean

Short Term Fuel Trim (STFT) is the real-time correction the ECU applies to its base fuel calculation, driven directly by oxygen sensor feedback. When the O2 sensors report lean, the ECU adds fuel and STFT goes positive — a reading of +15 percent means the computer is adding 15 percent more fuel than its base map called for. When the sensors report rich, the ECU subtracts fuel and STFT goes negative. Long Term Fuel Trim (LTFT) is the learned, slower-moving version — the ECU's memory of how much correction has consistently been needed.

Healthy trims live close to zero. Sustained values beyond about plus or minus 10 percent indicate a real problem that will eventually set a fuel trim code — P0171 for lean, P0172 for rich. The trims themselves rarely tell you the exact failed part, but the pattern of the trims across operating conditions almost always points at the system responsible.

Reading the patterns: idle vs. cruise

The single most useful trick in fuel trim diagnosis is comparing trim values at idle against trim values at cruise, because different failures leave different fingerprints.

A vacuum leak introduces a fixed amount of unmetered air. At idle, total airflow is low, so that fixed leak is a large percentage of the total — the ECU has to add a lot of fuel. At cruise, airflow is high and the same leak becomes a small percentage. So LTFT of +25 percent at idle dropping to +5 percent at cruise is the classic vacuum leak signature: big correction at low airflow, small correction at high airflow. Load-dependent trim behavior that shrinks with airflow is the fingerprint of a vacuum leak.

The opposite pattern — small positive trim at idle, larger positive trim at cruise, say +5 percent at idle growing to +12 percent at cruise — points to a MAF sensor under-reporting airflow, and under-reporting worse at high flow. Contamination on the sensing element, physical damage, or plain aging causes the MAF to report less air than the engine is actually ingesting, and the ECU adds fuel to feed the air it doesn't know about. MAF cleaning or replacement is the fix.

Strongly negative trims mean the ECU is pulling fuel out because something is adding fuel it didn't command. LTFT of -20 percent at cruise suggests excess fuel entering the system: a leaky injector dribbling fuel, high fuel pressure from a failed regulator, a saturated EVAP canister dumping hydrocarbons through the purge valve, or a contaminated MAF reading high — the ECU fuels for phantom air that isn't there.

Bank-to-bank splits: diagnostic gold

On a V-configuration engine, fuel trims on the two banks should track each other closely. When they split — say -8 percent on Bank 1 and +15 percent on Bank 2, or one bank at +18 percent while the other sits at -2 percent — you have a bank-specific problem, and that split does half your diagnosis for you. The bank with high positive trim is being under-fueled: look for a vacuum leak in that bank's intake runners or a bank-specific fuel delivery problem like a failing injector. The bank with negative trim is being over-fueled: think leaking injector on that bank, or an O2 sensor on that bank reading incorrectly. Critically, a normal-reading bank proves the shared fuel system upstream — pump, filter, common rail pressure — is functional, which eliminates a whole category of suspects.

Remember the convention: Bank 1 is the bank containing cylinder 1, Bank 2 is the opposite bank. But always verify the manufacturer's bank designations before condemning a part — getting the banks backwards means replacing parts on the wrong side of the engine.

When trims lie, swing, or miss the event

Fuel trims should move smoothly. Trims that swing rapidly between positive and negative indicate an unstable input, not an unstable engine: a failing O2 sensor sending erratic voltage is the most common cause, but a bad ground connection corrupting voltage references or a PCM problem can do the same. Scan tool history and a scope on the O2 waveform separate these.

Also understand what trims can't see. Fuel trims are averages over time. A brief transient event — a lean stumble on tip-in acceleration, for instance — can be completely invisible in trims that read a healthy plus or minus 3 percent, because the event is over before it moves the average. When the customer feels something the trims don't show, move to scope-level tools: watch the injector pulse width on a lab scope, or put a wideband O2 on the exhaust and watch actual air-fuel ratio during the acceleration event. Throttle body issues and momentary enrichment failures live in these transients.

MAF, MAP, TPS, crank position, and coolant temperature sensors feed the data every fueling and timing decision is built on — and each one has a known-good signature and a proper test procedure. Learn to verify a sensor instead of guessing at it.

Airflow and pressure: MAF and MAP

The mass airflow sensor reports airflow in grams per second, and the fastest sanity check is comparing its reading to expected values. Two rules of thumb: at idle, expect roughly 1 gram per second per 1,000 RPM on many engines, and idle MAF in g/s should roughly equal engine displacement in liters — a 2.0L engine should show around 2 g/s at idle. Compare readings at several RPM points against manufacturer data or a known-good reference vehicle. Deviations point to contamination on the sensing element (address with MAF-specific cleaner), physical damage, or wiring problems.

When a P0102 (MAF circuit low input) sets, work the sequence cheapest-first: check the MAF signal on the scan tool for plausible values, verify power and ground at the MAF connector, clean the element if it's dirty — a contaminated hot-wire under-reads — and check for intake air leaks between the MAF and the throttle body. Air entering through a leak downstream of the MAF is unmetered, which lowers the MAF's apparent reading relative to what the engine actually ingests. Only after all of that do you consider replacement.

The MAP sensor reads manifold absolute pressure and can be tested without a scan tool. Backprobe the signal wire with a DMM: at key-on-engine-off there's no vacuum, so it should read atmospheric pressure — around 4.5 volts or about 100 kPa. At idle, manifold vacuum is high, so the signal should drop significantly. The definitive test is applying a hand vacuum pump directly to the sensor and watching the signal track smoothly with the vacuum you apply. Note that some MAP sensors output a frequency instead of a voltage; the principle is the same.

Position and temperature: TPS, CKP, and coolant temp

A throttle position sensor should deliver a smooth, linear voltage sweep as the throttle moves from closed to wide open. Watch it on a scan tool (or better, a scope) while opening the throttle slowly: any sudden dropout to zero volts or spike indicates worn resistive tracks in a potentiometer-type sensor. Modern Hall-effect and other non-contact TPS designs fail differently — usually a sudden dead output or a stuck value rather than a scratchy sweep — but the same slow-sweep test reveals it.

The crankshaft position sensor is the master timing reference: the ECU cannot fire injectors or ignition coils without knowing crank position. That makes the failure symptoms unmistakable — complete CKP failure means the engine cranks but will not start; intermittent failure means unexpected stalling, often with restart after cooldown. Test it by scoping the signal while cranking: no signal, or a noisy and irregular signal, means a failed sensor or a damaged tone ring.

Coolant temperature sensors are thermistors — resistance changes with temperature — and the ECU converts resistance to a temperature reading. This gives you a built-in fault signature: an open circuit (broken signal wire or a sensor failed open) reads as maximum resistance, which the ECU displays as -40 degrees F, the bottom of its scale. So a scan tool showing -40 on a fully warmed engine at 180 degrees means an open circuit, not a cold engine. A shorted circuit reads the opposite extreme — maximum temperature. Either way it needs repair, because the ECU makes cold-enrichment and other fueling decisions from this signal.

Catching the intermittent: data logging

The hardest sensor faults are the ones that leave no codes. For an intermittent stall with a clean code scan, the answer is not more parking-lot testing — it's capturing data during the event. Use a scan tool with recording or graphing capability (or a dedicated data logger) set up to log RPM, sensor values, fuel pressure, and ignition timing while the customer or you drive until the stall occurs. Then review the capture: what changed at the moment of failure? Did RPM collapse instantly (think crank sensor or ignition), did a sensor value spike, did fuel pressure fall away? The recorded data reveals the failure pattern that a static test never will.

When communication faults appear, techs who don't understand the network replace modules at random; techs who do can isolate a bus fault to a wire, a terminator, or a single dead node with a meter and a scope.

How CAN works

CAN bus — Controller Area Network — is a serial communication network that lets multiple modules (ECU, transmission controller, ABS, body control module, and dozens more) share data over just two twisted wires, CAN-High and CAN-Low. Each module broadcasts its messages onto the shared bus — RPM, coolant temperature, vehicle speed, switch states — and every module sees every message, responding to the ones relevant to it. This design eliminates the need for a dedicated wire for every signal between every pair of modules. The trade-off: because the infrastructure is shared, a bus fault tends to affect multiple modules simultaneously, which is itself a diagnostic clue.

CAN uses differential signaling: CAN-H and CAN-L carry mirror-image versions of the same data, which lets receivers reject noise that hits both wires equally. Modern vehicles actually run several CAN networks at different speeds — a high-speed bus for critical systems like powertrain and ABS, a low-speed bus for body electronics, and often separate chassis and infotainment buses. A gateway module bridges these networks, translating and filtering messages between them and providing security and rate limiting between buses. Your scan tool usually connects through that gateway, which affects what data you can reach and how.

Diagnosing communication faults

The pattern of failure tells you where to look. When multiple modules show communication errors at once, the problem is almost certainly shared infrastructure, not several modules dying together. Check power and ground at the affected modules first — supply problems are common and cheap to find. Then move to the bus itself: inspect CAN wiring for shorts and breaks, and measure the terminating resistors. With the modules powered down, resistance between CAN-H and CAN-L should read approximately 60 ohms — that's two 120-ohm terminating resistors at the ends of the bus in parallel. A missing or failed terminator corrupts communication across the entire network. Don't forget gateway modules, since a dead gateway can silence a whole bus segment.

When a single module shows no communication, the sequence is: verify battery and ignition power and ground at the module's connector, then confirm CAN signals are actually reaching the module. If power, ground, and bus signals are all present and the module still won't respond to the scan tool, the module itself is likely internally failed — but only condemn it after confirming all of its supporting infrastructure. Modules are expensive and often require programming; wiring is cheap.

For intermittent or subtle bus faults, the lab scope is the tool. Put one channel on CAN-H and one on CAN-L: a healthy bus shows mirror-image waveforms centered around 2.5 volts, with CAN-H peaking near 3.5 volts as CAN-L drops to about 1.5 volts. A wire shorted to power or ground shows as a stuck, flat line. Damage to the twisted pair shows as asymmetric noise between the channels. Corrupted, malformed patterns point to a failing transceiver somewhere on the bus.

Aftermarket accessories and the bus

The CAN bus is a precision-engineered network, and careless accessory installation is a common source of mystery network faults. Improperly wired or unlicensed devices can flood the bus with messages, change the bus impedance, or interfere with the terminating resistors — all of which produce network faults that are miserable to trace later. Use manufacturer-approved connection methods: add-a-fuse taps and designated accessory power points for anything that just needs power, and certified aftermarket CAN gateway interfaces designed for the specific vehicle when a device genuinely needs to talk on the network. Never splice directly into CAN wires to power an accessory.