What matters
- Always compare actual values with their targets and align events by RPM and time.
- A boost error, timing correction, or fuel-pressure dip needs context from related channels.
- A general analyzer can screen a pull, but the platform calibrator remains the authority.
Begin with the test conditions
Before judging a graph, identify the fuel, calibration, gear, road or dyno, ambient conditions, starting temperatures, and whether traction control or a shift interrupted the pull. Comparing two logs is useful only when the conditions and channel definitions are reasonably consistent.
Use a channel stack, not a single line
Start with RPM, accelerator position, throttle, and gear or vehicle speed to identify a clean wide-open-throttle section. Then compare boost actual with boost target, lambda actual with lambda target, and fuel pressure actual with fuel pressure target. Add per-cylinder timing corrections, intake-air temperature, and fuel trims for context. COBB’s logging documentation similarly describes target-versus-actual comparisons for boost and rail pressure and explains how barometric pressure changes the work required of the turbocharger.
At 3,500 RPM, target boost is 18 psi while actual briefly reaches 21 psi. Throttle then closes and actual boost falls below target near 5,500 RPM. That pattern can be consistent with an overshoot followed by torque intervention, but the log must be checked for traction control, wastegate command, gear change, and sensor conventions before assigning a cause.
Timing corrections require context
A correction is not automatically proof of destructive knock, and the meaning differs by platform. Look for depth, repetition, cylinder concentration, load, temperature, and whether the event repeats with verified fuel. A single transition event deserves different treatment from repeated multi-cylinder corrections under sustained load. Stop testing and consult the tuner when the pattern is severe, repeatable, or accompanied by lean mixture, pressure loss, misfire, or abnormal sound.
Fuel pressure and mixture belong together
If rail pressure falls away from target while lambda also moves leaner than commanded, the combined evidence is more serious than either line alone. If pressure is stable but trims climb, investigate the airflow model, leaks, sensor scaling, and fuel composition. Never apply another platform’s “normal” threshold without confirming the units, sensor location, and ECU strategy.
Sources and methodology
Calculations are shown so the assumptions can be checked. Safety-critical fitment, fueling, and calibration decisions must be verified against current manufacturer documentation and the professional responsible for the work.