Diagnostics

On a mobile call you get one visit, the customer watching, and whatever the vehicle decides to do while you are standing there. Field diagnostics is about extracting the maximum information in minimum time — and refusing to guess with the customer's money.

Start with the Scanner, and Buy the Right One

For a cranks-but-won't-start complaint, the single tool that tells you the most in the shortest time is the OBD-II scanner. In one connection you get stored codes, live data — fuel pressure sensor readings, RPM signal, cam and crank correlation — and freeze frame data. That combination rapidly sorts the problem into fuel, spark, or timing territory without removing a single component. Compression testers, vacuum gauges, and timing lights all have their place, but they each test one system; the scanner triages all of them at once.

When you buy that scanner, the feature worth paying for is bidirectional control — the ability to command components on demand. Being able to activate a fuel pump, fire an injector, run a cooling fan, or cycle an ABS actuator from the tool proves whether the component works before you order parts. Reading codes, viewing live data, and Bluetooth convenience are all standard fare; bidirectional control is frequently the difference between a confirmed diagnosis and an educated guess, and on mobile work a wrong guess means a second trip and an awkward conversation.

Codes Are Clues, Not Diagnoses

A trouble code tells you what the computer noticed, not what is broken. P0420 (catalyst efficiency below threshold) is the classic trap: it can be triggered by lazy upstream O2 sensors, exhaust leaks, engine misfires, or fuel system problems — not just a dead catalytic converter. Before you quote a catalyst replacement north of 600 dollars, check the upstream O2 sensor signals, look at fuel trim data, and rule out the cheaper causes. Sell the cat without doing that homework and the light comes back on in a week, along with your credibility.

P0171 (system too lean, bank 1) follows the same logic. The number one real-world cause of lean codes on gasoline engines is a vacuum leak — unmetered air entering the intake downstream of the MAF sensor. Because that air never passed the MAF, the ECU does not know it exists and cannot add fuel to match it, so the mixture runs lean. Smoke test the intake or check for vacuum leaks before condemning O2 sensors, MAP sensors, or anything else.

A stored-but-not-current code like P0300 (random/multiple misfire) is not something to ignore or to answer with a blanket tune-up. Pull the freeze frame: it records the RPM, engine load, coolant temperature, and throttle position at the moment the code set. Duplicate those conditions during your test drive or testing and you can catch the misfire live and identify which cylinder is acting up. The freeze frame is the vehicle telling you exactly where to look — use it.

Isolating a Weak Cylinder

When an engine runs rough and a cylinder balance test on the scan tool shows one cylinder — say cylinder 3 — contributing noticeably less than the others (15 percent down, for example), you have localized the problem but not identified it. A weak cylinder has three possible causes: fuel (injector), spark (coil or plug), or mechanical (compression). The next move is a compression test on that cylinder, because it cleanly separates mechanical failure from ignition and fuel causes. Good compression means you can focus on spark and fuel; low compression means no plug, coil, or injector will fix it, and the customer needs a very different conversation.

For the fuel side, there is a fast non-invasive check: put a mechanical stethoscope — or a long screwdriver used as one — against the injector body while the engine runs. A working injector clicks with every pulse. No click means the injector is not receiving a signal or not operating, and you learned it without removing anything. It is the fastest injector test in the field.

Electrical Testing: Measure, Don't Assume

Fuel pressure is a case study in trusting direct measurement. Newer vehicles display fuel pressure sensor data on the scan tool, but a sensor can lie — and when the sensor itself is a suspect, its own data proves nothing. A mechanical fuel pressure gauge connected at the schrader test port or teed in-line gives you the actual pressure. When sensor data and mechanical readings can disagree, the gauge wins.

The same principle applies to checking power at a component. A test light only tells you voltage is present or absent; it says nothing about quality. A corroded wire with high resistance will happily light a test light while starving the actual component under load. Use a digital multimeter to measure the real voltage, and whenever possible back-probe the connector with the circuit under load — that is when resistance problems reveal themselves.

Voltage drop testing is the tool for finding that hidden resistance, and the spec matters: voltage drop across a battery cable under load should be under 0.2V. A reading like 0.6V means excessive resistance in the cable or its connections — corrosion, a loose terminal, a degraded crimp — and it needs cleaning or replacement. Degraded cables are the root cause behind a huge share of intermittent starting and charging complaints that get misdiagnosed as bad batteries and bad alternators. Measure the drop before you condemn the big components.

Intermittent Faults: Capture, Don't Guess

Intermittent problems are where mobile mechanics either build a reputation or wreck one. The temptation is to replace whatever commonly fails and hope; resist it. Parts-swapping is expensive, proves nothing, and when the fault returns the customer has paid for parts that fixed nothing. The professional approach to any intermittent electrical fault is to capture data during the failure event — a data logger or scan tool freeze-frame capture while duplicating the fault gives you evidence instead of guesses.

Setup matters. For an intermittent no-start, the effective approach is triggered logging with a pre-trigger buffer: log the crank and cam signals, fuel pressure, and RPM, triggered on the ignition-on event, with the buffer preserving the seconds leading into the failure. That captures the conditions before and during the fault — the data that actually diagnoses it. Logging everything constantly at maximum rate just fills storage, and starting a log after the fault occurs misses the failure moment entirely.

Then there is the awkward field reality: you arrive for a car that sometimes will not start, and it starts perfectly while you watch. The professional response is honesty. Explain that intermittent no-start diagnosis requires duplicating the fault, and offer a systematic plan — install a data logger for several days, or arrange conditions and timing so you can be present when the fault happens. Never sell parts as guesses to justify the trip fee. Customers respect a mechanic who says exactly what it takes to catch the problem far more than one who charges them for a maybe.