DIAGNOSTIC THEORY
Back to Explainers

What your scan tool can and cannot prove about a crank request

A scan tool only knows what it sampled while it was sampling. The crank request lives and dies between samples — read it during the attempt, not after it.

Conceptual definition

The "crank request" is the module's digital record that you asked for cranking — the ignition switch or push-button input translated into a signal the starting circuit acts on. Your scan tool reports that record as data PIDs. But a PID is a photograph, not a continuous witness: the tool samples each value on a schedule, and between samples it shows you nothing. Everything the tool can prove about a crank request is bounded by what it sampled, when it sampled it, and whether the value ever actually changed state.

This is a discipline explainer, not a parts explainer. It teaches how to read the tool's crank-request PIDs the way a witness statement is read — checking who was actually watching, and when.

The PIDs that carry the request

Not every PID with "crank" or "starter" in the name means what you think it means. Four carry the request story, and each one makes a different claim:

Request vs command vs execution: three separate claims

A crank attempt is three claims stacked on top of each other, and the tool can watch all three — but it must not blur them:

Request without command means the module refused — inhibit logic, anti-theft, a missing enable. Command without execution means the request chain is intact and the failure is downstream — relay, wiring, starter, power, ground. Execution with a weak crank is a different question entirely, and the tool's job there is over. The three claims separate the wiring side from the module-decision side from the mechanical side, and mixing them up is how no-cranks get misdiagnosed.

Sample latency vs the START pulse

A keyed START pulse is brief — a flick of the wrist, a second or two of the switch held in START. A scan tool sampling a dozen PIDs over a busy data bus refreshes each one far slower than you assume, often a few times a second at best, slower when the module list is long. A brief START pulse can come and go entirely between samples of the Crank Request PID.

This is the central trap: you attempt a start, the engine doesn't crank, you look at the tool afterward and the Crank Request PID reads NO. That read proves nothing about the attempt. It proves the tool wasn't watching when it happened — or that it watched and missed it. An after-the-fact read of a momentary signal is not evidence of anything. The only read that carries weight is a read captured DURING the attempt, with your eyes on the PID while the switch is held or the button is pressed.

Discipline: set up the data screen first, get the PID visible and refreshing, THEN attempt the start while watching. A read you recorded after the attempt is a read of whatever state the vehicle fell back into — the request state is gone.

Frozen-PID discipline: a value that never changes is not a measurement

Before you trust any PID, prove the tool is actually receiving it. A scan tool will happily display a default, substituted, or stale value with the same confidence it displays a live one. The check costs five seconds: note the PID's state, change the real condition (press the button, turn the key to START), and confirm the PID changes with it and changes back when you release.

If the Crank Request PID reads NO while you're holding START and never flickers, that is not proof of a missing request — it might be a dead data path, a module that doesn't broadcast that PID on this vehicle, or a tool that maps that parameter name to nothing. A PID that never changes state is decoration, not a measurement. Prove liveness on every PID you intend to read before you read it for evidence.

Same caution for Power Mode: if it reports RUN while you're in START, or never reports START at all on a push-button vehicle, check the PID's behavior across modes (OFF/ACC/RUN) before concluding the request never arrived. Some vehicles simply don't broadcast a START power-mode state — the absence of the state in the tool is not the absence of the request in the vehicle.

Push-button vs keyed reporting differences

Keyed and push-button systems report the request differently, and the difference matters:

On push-button vehicles, "no crank request" almost always means the conversation broke down somewhere — the button input never got translated into a request — not that the button itself failed. The tool's Power Mode and request PIDs are the transcript of that conversation; read them for the sequence, not just the final state.

What it is not

How SpeedNeeds uses it

The scan tool's crank-request evidence is read with the same discipline as any other sampled instrument: liveness proved first, the read taken during the attempt, and the three claims — request, command, execution — kept separate. A request seen live during the attempt is a proven input; a command absent under a proven request names the module's decision boundary; execution confirmed by RPM closes the chain. After-the-fact reads are treated as untested, not negative. And where the PIDs disagree with physical reality — request shown but relay never pulled, command shown but no motion — the tool's transcript is used to locate the break, never to override what the meters and ears already proved.

Caution and edge cases

Closing clarity

The scan tool is a competent witness and a terrible narrator: it reports exactly what it sampled, and it says nothing about what it missed. Read the request PIDs live during the attempt, prove they change before you trust them, and keep request, command, and execution as three separate claims. A no-crank with a proven request and an absent command is a module decision — find out why it refused. A no-crank with a proven command and no RPM is a downstream failure — go measure power and ground. The tool gives you the map; the meters prove the ground.

Copyright 2026 SpeedNeeds.net. All rights reserved.