Boost logs · a measurement guide
How to read requested vs actual boost
The ECU asks for a pressure. A sensor reports what happened. The difference is useful evidence, but it does not identify a failed turbo, wastegate, hose or sensor on its own. This guide shows how to make that comparison without turning an observation into a guess.
1 — Capture the channels that answer the question
At minimum, keep time, engine speed (RPM), requested boost and actual boost in the same log. Throttle or accelerator position helps identify a sustained load region; a barometric-pressure channel helps reconcile absolute and gauge readings. Keep the original export so its headers and units remain visible. Collect a log in a controlled, safe setting or ask a qualified workshop to do it.
If the ECU does not expose a requested-boost channel, a single actual-pressure trace cannot tell you whether it missed the target. Our real-data example demonstrates that refusal: the log contains a genuine pull but no usable target.
2 — Put target and actual on the same pressure scale
Boost logs may use bar, kPa, mbar or psi. Some channels report absolute pressure; others report gauge pressure, which has atmospheric pressure removed. Comparing an absolute target directly with a gauge actual creates an artificial gap of roughly one atmosphere. Read the channel headers and any logged barometric value before interpreting the numbers.
If both traces use the same unit and datum, target minus actual is the shortfall even when the exact absolute pressure is uncertain. If their units or datum differ, convert them first. When a file does not establish the datum, the analyser prints that uncertainty instead of silently treating an inferred value as a measurement.
3 — Compare a sustained commanded region
A brief tip-in, a gear change or a lifted pedal can create a temporary separation that is not the same as a sustained underboost. Compare the traces only where the ECU is actually requesting boost and the log contains enough samples and RPM range to describe a pull. The analyser finds that region using pedal, load or boost request where available; when it cannot find one, it declines to judge.
Record the conditions alongside the trace: RPM span, gear or load, and whether a related DTC was stored. A log cannot replace the vehicle-specific diagnostic procedure or confirm that a repair worked without a comparable after-log.
4 — Read the shape before naming a cause
| Trace shape | What the log supports | What remains unproven |
|---|---|---|
| Persistent undertrack | Actual stays below request through the commanded region. | Which part caused the shortfall. |
| Slow response | Actual starts below request, then catches up. | Whether the delay is mechanical, control-related or expected for this calibration. |
| Build then collapse | Actual reaches request, then falls while request remains high. | Why delivery stopped or whether another protection intervened. |
| Overshoot | Actual rises above request before coming back. | Which control element produced the excess. |
| Erratic signal | Actual repeatedly moves away from its local trend. | Whether the movement is physical or a measurement problem. |
| Insufficient data | No defensible requested-vs-actual verdict. | Any boost fault from that file alone. |
These are descriptions of recorded behaviour, not a ranked list of repairs. The next test depends on the engine's boost-control architecture and companion codes. Start with P0299 underboost, P0234 overboost or P2263 boost-system performance for the diagnostic order.
Example shape: actual builds but plateaus around 0.9 bar and never reaches the 1.40 bar request — the gap is the fault.
The trace above is illustrative synthetic data. It shows the meaning of a persistent gap; it does not describe a customer vehicle.
Read your own log
The free analyser runs in your browser and reports its assumptions, the region it judged and the gap it measured. Your file stays on your machine.
Open the boost log analyser →For how we separate observations, evidence and sign-off, read the methodology.
This guide explains how to read a measurement, not how to authorise a repair on a particular vehicle. It has no workshop sign-off. Confirm any diagnosis against the manufacturer's procedure and measured evidence.