GM P2138 and Reduced Engine Power After Rain: Preserve the Wet-Event Evidence
P2138 After Rain
GM code P2138 means two accelerator-pedal position signals disagree; it does not automatically mean the pedal assembly failed. On an applicable 2018-or-prior GM passenger car or light-duty truck, timing after rain, a car wash, wet carpet, windshield work, or cowl leakage matters because bulletin 07-06-04-019F identifies water intrusion in an instrument-panel-to-body harness area as one path to investigate.
Once the correct diagram and circuits are known, a meter with a compact waveform view can compare the paired signals without turning a connector into a parts guess. The AUTOOL DM301 automotive multimeter/oscilloscope is one measurement route. Its powered functions must not be applied to sensor, module, network, SRS, or unknown circuits, and a trace cannot identify the source of water or condemn the pedal by itself.
Treat the vehicle like a wet-event incident file. What got wet, when the message appeared, which signal disagreed, and what dried out are separate pages.
Quick answer: P2138 means the paired accelerator signals disagree, not that the pedal is proven bad. Preserve wet-event evidence and test the exact circuit path.
In this guide
Open an incident file before it dries
If Reduced Engine Power appears in traffic, move out of the lane only if the vehicle responds safely. Do not demand full acceleration. If throttle response is severely limited, the engine runs poorly, or the vehicle cannot maintain a safe speed, stop and arrange recovery.
Before cycling the key, photograph the message and lamps. Record rain intensity, puddle exposure, car-wash type, parked direction, HVAC mode, wet shoes, door/window use, and elapsed time since the water event. Check front carpets, sill, kick-panel area, dash underside, cowl, and visible connectors without pulling trim around airbags.
Do not dry, disconnect, spray, or reseat a suspect connector before scanning and photographing it. The first evidence may disappear quickly, while corrosion or terminal damage remains hidden.
Save a complete system scan with P2138 status, freeze frame, failure mileage, voltage, and every related 5-volt-reference, throttle, body, communication, or water-sensitive module code. Current versus history status helps place events in order.
Prove the vehicle is inside the bulletin fence
GM bulletin 07-06-04-019F covers 2018 and prior GM passenger cars and light-duty trucks, but exact vehicle architecture, connector location, and service method still control. Verify VIN, model, build, engine, market, and the latest bulletin/service information.
The document’s water-intrusion route is not a statement that every pre-2019 GM with P2138 has a wet connector. If the event has no moisture correlation, if another code identifies a different reference circuit, or if the vehicle uses another harness layout, return to the appropriate service-manual tree.
Ask about windshield replacement, sunroof-drain work, cowl service, body repair, door-seal leaks, HVAC drain issues, accessory wiring, pedal replacement, and floor flooding. A leak repair performed months earlier can leave corrosion that fails only when humidity rises.
Compare both pedal signals as a pair
Accelerator pedal position uses redundant signals so the controller can detect disagreement. Inspect live APP1 and APP2 data from rest through a slow, smooth pedal sweep with the engine off or in the exact safe state specified by GM. Both should change smoothly and maintain the relationship defined for that vehicle.
Record percentage or voltage at several pedal positions, but do not impose one generic ratio on every GM system. Look for jumps, dropouts, flat spots, disagreement during connector movement, and changes after humidity/temperature shifts. Compare scan data with direct circuit measurements only when the correct terminals and high-impedance method are known.
If both signals drop together, consider shared supply, ground, connector, or controller paths. If one changes and the other remains stable, follow the individual circuit. If scan data disagrees but both direct signals remain plausible, network/controller interpretation and test setup move up the list.
Trace where water could cross the circuit
Water follows gravity, harness tape, capillary action, body seams, and connector cavities. Map the highest fresh wet point and the lowest collected water. A damp connector does not prove the leak originated there.
Inspect cowl drains, windshield sealing, body seams, HVAC drainage, door membrane, grommets, and prior holes or accessory pass-throughs. Use approved leak-detection methods that do not flood modules or create a new intrusion. Avoid high-pressure water aimed at seals.
At the connector named by the exact GM procedure, document moisture, residue, terminal tint, seal damage, fretting, looseness, and wire tension. Repair terminals and seals using approved parts and crimp methods. Cleaning corrosion without stopping the water source creates a temporary result.
Inspect and measure without damaging terminals
Disconnect the battery only when the service procedure requires it and after preserving scan data, security, memory, and module-sleep considerations. Work around SRS areas only with the specified disable time and trained process.
Use breakout leads or approved back-probing tools. An oversized meter probe can spread a female terminal and create another intermittent open. Never puncture insulation unless the repair method permits and seals the conductor afterward.
Perform voltage-drop or continuity checks under the condition that reproduces the failure. Open-circuit continuity can look perfect when a corroded strand or weak terminal cannot carry the actual signal/reference current. Conversely, resistance testing on a live module circuit can damage the meter or module.
Do not use a powered probe on APP circuits. The controller’s 5-volt reference and sensor returns are not ordinary 12-volt loads.
Keep P2138 from becoming a pedal order
The pedal assembly remains a valid branch when its redundant tracks fail the specified sweep or connector test. So do harnesses, terminals, reference/ground circuits, water intrusion, shared 5-volt loads, controller inputs, and software. The code tells you why the controller limited power—not which box to buy.
A pedal that works after key cycling is not cleared. A connector that looks dry after an hour is not cleared. A repair that restores signal correlation but leaves wet carpet is not finished.
Give the shop the incident file: VIN/build, event weather, exact message, current/history codes and freeze data, APP1/APP2 graph, wet-area map, recent glass/body/accessory work, connector photographs, circuit measurements, leak source, terminal repair, and post-repair results.
Verify through a dry cycle and a controlled recheck
After stopping the leak and repairing the proven circuit, dry affected cavities by the approved method and confirm seals, drains, harness routing, and carpet/insulation condition. Clear only resolved faults.
Repeat a slow APP sweep and verify both signals remain smooth and correctly correlated through connector/harness movement allowed by the procedure. Perform a safe road test through normal throttle ranges, then re-scan after a sleep/restart.
Where a controlled water test is authorized, recreate the source—not by flooding the electrical area—and confirm no moisture arrives. The case closes when P2138 and Reduced Engine Power do not return, the signal pair remains stable, the cabin stays dry, and the recorded leak/circuit repair explains the original rain link.
Comments 0
Questions, fixes, and real-world diagnostic notes from readers.
Be the first to share a diagnostic result or ask a follow-up question.