Probe placement is the single most common cause of wrong readings on a 5-gas analyzer. If your CO + CO₂ total at idle is below 12 %vol, don't start diagnosing the engine — fix the sampling first.
The 12% rule
Combustion of hydrocarbon fuel produces CO and CO₂ in predictable proportions. At stoichiometric idle, CO + CO₂ sits in the 12–16% band. When ambient air sneaks into the sample the dilution pulls that sum down.
// check first
If CO + CO₂ < 12%, you are probably either too shallow or reading through an exhaust leak. Investigate before trusting lambda.
Depth and angle
- Minimum depth: 30 cm into the tailpipe (further on large-bore systems).
- Avoid direct exhaust flow: angle the probe to the side to prevent sensor damage and fluctuating readings.
- Twin-pipe vehicles: test one at a time; confirm both read identically (difference can indicate bank/cylinder imbalance).
- Short tailpipes / sports exhausts: fit a temporary extension so the probe tip is past the pipe mouth.
Leak-before-probe pitfalls
An exhaust leak upstream of the probe adds atmospheric oxygen and drops CO₂. You'll see O₂ climb, λ calculate lean, and CO₂ suppressed — a pattern that mimics a vacuum leak. Confirm by running the engine at 2500 rpm: a real leak usually reads closer to stoichiometric at elevated rpm (higher exhaust pressure seals the leak or overwhelms the ambient ingress).
Water condensation
Cold vehicles produce steam and water at the tailpipe that can reach the analyzer water trap. Warm the engine fully, drain the trap and watch for droplets in the line.
Checklist
- Engine fully warmed (coolant > 80°C).
- Probe ≥ 30 cm deep, not facing flow.
- CO + CO₂ at idle ≥ 12%.
- No water in the trap, no flashing sample-low error.
- Retest if exhaust tone changes mid-sample.