Is it safe to keep driving?
A steady shake with the engine light off or steadily on is usually safe for a short drive to a workshop. A flashing engine light while it shakes is not — that means raw fuel is reaching the catalytic converter and you can destroy it in a few kilometres. Pull over and stop driving.
General guidance, not a diagnosis of your specific vehicle. If you are unsure, do not drive it and have it inspected.
What the shake is telling you
At idle the engine is producing just enough torque to keep itself turning, so a single weak cylinder is a large fraction of the total. That is why a misfire you cannot feel at 100 km/h is obvious at a stop light. The same logic applies to air leaks: an unmetered leak that is trivial next to the airflow at cruise is a large proportion of the tiny airflow at idle, so the mixture goes lean and the idle hunts.
Separating a misfire from a mixture problem
A misfire is usually rhythmic — you can feel a repeating stumble. A lean idle is more of a hunt or a sag, with the revs wandering up and down as the ECU chases the target. Watching live data settles it quickly: misfire counters climb on specific cylinders, whereas a vacuum leak shows as high positive fuel trim that falls back toward zero as you raise the revs.
The vacuum-leak test that needs no tools
Note the fuel trim at idle, then hold the engine at about 2,500 rpm and watch it again. If short- and long-term trim were strongly positive at idle and drop back toward zero at higher revs, the engine is drawing unmetered air — a split intake hose, a perished gasket, a leaking PCV line. If trim stays high across the whole rev range, look instead at fuel delivery or the mass airflow sensor.
Mounts, not the engine
If the engine runs cleanly, stores no codes and the misfire counters stay at zero, but the whole car shakes when you put it in drive with your foot on the brake, suspect collapsed engine or transmission mounts. A failing mount transmits normal combustion vibration into the body instead of absorbing it. The scan tool will show nothing, which is itself the clue.
What to watch in live data
These are the readings that separate the likely causes. OBDAssistant streams them from an ELM327 adapter while the engine is running, which is what turns a list of possibilities into a diagnosis.
| Reading | What it tells you |
|---|---|
| Misfire counters per cylinder | All zero rules a misfire out. One cylinder climbing points at its coil, plug or injector; several climbing together usually means fuel, vacuum or timing, not ignition. |
| Short-term and long-term fuel trim | Strongly positive at idle and falling at 2,500 rpm points to a vacuum leak. Strongly positive everywhere points to fuel supply or airflow metering. |
| Engine coolant temperature | A cold engine idles rough legitimately. Confirm the engine actually reached operating temperature before diagnosing anything. |
| Calculated load and MAF g/s | An idling engine that reports implausible airflow is being lied to by its own sensor. |
Trouble codes that match this symptom
If the car has already been scanned, these are the codes most often behind it. Each page covers the definition, likely causes, diagnostic steps and repair information.
- P0300 Random or Multiple Cylinder Misfire Detected
- P0301 Cylinder 1 Misfire Detected
- P0302 Cylinder 2 Misfire Detected
- P0303 Cylinder 3 Misfire Detected
- P0304 Cylinder 4 Misfire Detected
- P0171 System Too Lean (Bank 1)
- P0101 Mass Air Flow (MAF) Circuit Range/Performance Problem
- P0505 Idle Control System Malfunction
- P0172 System Too Rich (Bank 1)
Background reading
- Fuel Trim Mastery: OBD2 Diagnostic Guide
- ECU Knock Logic: Ignition Timing & Detonation Protection
Questions
Can a dirty throttle body cause a shaking idle?
Yes. Carbon builds up on the throttle plate and bore, and the idle air control strategy loses the small, precise airflow it is trying to meter. It usually shows as an unstable or low idle that improves briefly when you blip the throttle, often alongside an idle control code.
Why does it shake only when the air conditioning is on?
The compressor adds load the moment it engages. An engine that is already marginal — a weak cylinder, a small vacuum leak, a tired idle control valve — cannot absorb it, so the shake appears with the A/C and disappears without it. The underlying fault is still there.
The shake went away on its own. Do I still need to look at it?
If a code was stored, yes. Intermittent misfires that clear themselves usually come back, and the stored freeze-frame data captured the conditions at the time, which is the most useful diagnostic evidence you will get.