P24C4: An EGR bypass learning limit is not the normal travel limit

P24C4 concerns the learned electrical positions of the EGR cooler-bypass valve. The monitor asks whether the endpoint feedback falls inside expected learning windows, rather than simply checking whether a commanded move was completed. A valve that moves can still produce an unacceptable learned endpoint.

Two endpoints provide different evidence

The cooling-position window is bounded by 16% and 24% of the 5 V reference. The bypass endpoint uses a separate, wider window of 60.9% to 91.4%. These are raw-voltage learning criteria, not interchangeable percentages of valve opening.

The check follows successful learning in fully closed and fully open positions at key-off. The source also specifies coolant-temperature eligibility, a valid temperature signal, and absence of relevant actuator, position-sensor and position-deviation faults. Without a completed eligible learning event, the endpoint comparison cannot be assumed to have run.

Preserve the learning context

The row reports no additional debounce after an eligible out-of-window result. Its two-trip classification should not be mistaken for a command to repeat an arbitrary actuator cycle twice. The end-of-trip event and learning prerequisites still matter.

Useful evidence includes which endpoint exceeded its window, the learned raw value and whether learning completed normally. A normal reading in mid-travel does not validate an endpoint offset. The table identifies a learning-range problem but does not specify a replacement part, adjustment or universal relearn procedure.

Source and applicability

2025 GM 25OBDG04A documented monitor only. Vehicle, engine and calibration applicability must be confirmed from the matching service information; no model fitment is inferred. These criteria explain controller logic, not a complete repair procedure. 25OBDG04A, PDF page 1225.