What matters
- Actual values need commanded targets and operating context.
- Short transitions and sustained deviations should not be treated the same.
- A general tool can flag rows but cannot assign a component failure from correlation alone.
Why this question matters
A datalog is a time-aligned record of channels, not a diagnosis. Useful interpretation begins with channel definitions, targets, operating context, and repeatability.
Build a defensible log-review order around commanded targets, operating state, units, and sustained differences instead of universal thresholds. The goal is to make the assumptions visible so the result can be checked, repeated, and updated when vehicle technology or official guidance changes.
What the evidence supports
Platform logging guides commonly expose target and actual channels for boost and fuel pressure because control error is more meaningful than an isolated absolute value. The channel name, units, scaling, sampling, and ECU strategy must still be confirmed.
Start with a matched pair, not two similar names
A valid comparison requires the commanded and measured channels to describe the same quantity, reference, units, and location. COBB’s Toyota guide, for example, identifies its boost actual and target pair as absolute throttle-inlet pressure after the intercooler but before the throttle body; manifold absolute pressure is a different channel at a different location. Comparing either one with a generic “boost” target can create an error that exists only in the spreadsheet.
The same rule applies to fuel pressure and mixture. Confirm high-side versus low-side pressure, gauge versus absolute pressure, lambda versus gasoline-scaled AFR, and final target versus an intermediate table value. Preserve the original headers and logger documentation in the review record so a later software update does not silently change the interpretation.
Ask whether the target is active at that moment
A target can exist in a log while the controller is outside the state that tries to achieve it. Haltech documents closed-loop boost conditions such as a control-point offset, a maximum pressure derivative, throttle requirements, and activation delays. During spool, a shift, traction intervention, throttle lift, protection, or an open-loop mode, the difference between target and actual may be expected rather than evidence of failed hardware.
Mark pedal, throttle, RPM, gear, control mode, limits, and status flags before measuring error. Evaluate a stable interval where the relevant controller is active, then separately describe entry, overshoot, settling, and exit behavior. A one-row maximum combines these phases and discards the state information needed to understand them.
Turn the difference into a testable question
Calculate error only after the channel pair is validated. Absolute error keeps the engineering units visible; percentage error can help compare operating levels but becomes misleading when the target is near zero. Record duration, RPM span, repetition, and direction instead of reporting only the worst point.
Then add the control and consequence channels. For boost, review throttle, wastegate or actuator command, manifold pressure, barometric pressure, and protection state. For rail pressure, review low-side supply, pump command, injection demand, voltage, and lambda target versus actual. The result should be a narrow next test—such as checking whether a sustained pressure shortfall repeats under the same demand—not a component verdict from correlation alone.
Twenty psi of boost can be exactly on an entered 20 psi target, 3 psi over a 17 psi target, or 5 psi under a 25 psi target. The absolute number is identical; the control-system meaning is not.
A repeatable workflow
- Mark a clean operating range with RPM, pedal, throttle, and gear.
- Confirm channel definitions and units.
- Overlay actual with commanded boost, lambda, and pressure where available.
- Inspect sustained differences and correlated control channels before forming a platform-specific question.
Where the shortcut breaks
Targets are not automatic safety limits. A commanded value can still be inappropriate, a sensor can be wrong, and a clean trace cannot certify engine health.
Sources and update method
TunerBench prefers government, standards-body, and component-manufacturer documentation. This guide is reviewed against the sources below and should be revisited when regulations, product data, or vehicle technology changes.