Some faults only exist while the engine is moving. The rock test makes the engine move on command — and watches the misfire counters for the answer.
The engine moves on its mounts. Torque twists it under load, and it rocks fore and aft when the transmission engages drive or reverse. Every wiring harness, connector, and ground strap attached to that engine flexes with it.
A terminal that is cracked, a pin that has backed out of its connector, or a wire chafed down to a few strands can make and break contact as the engine moves — an intermittent connection that never shows up with the engine sitting still in the bay. The fault exists only in motion.
The rock test reproduces that motion on command: foot on the brake, shift between drive and reverse, and watch the misfire counters while the engine rocks — with road speed at zero. If the misfire follows the movement, the fault is movement-correlated.
Worn mounts let the engine travel farther — a connection that survived for years starts breaking contact once the mounts sag.
Heat cycles expand and contract connectors a fraction of a millimeter thousands of times; eventually a marginal pin walks out.
Vibration plus movement frets terminals — microscopic wear that builds resistance at exactly the flex point.
The signature is what separates it: the fault follows movement, not load. It appears on tip-in, on gear engagement, or over bumps — and vanishes at steady cruise. A fault that tracks steady load is a different animal; a fault that tracks the rock is this one.
It is not a load problem. The engine rocks with zero road speed and near-zero load — which is exactly why the test can separate movement from load.
A rock test that moves the engine without moving the misfire counters proves the fault is not movement-correlated. Correlation needs the counters; a rock that 'felt wrong' without counter movement is not evidence of anything.
It is not a substitute for separating load from RPM first. Run the load-vs-RPM separation test before the rock test, so you know you are chasing movement and not a load problem wearing a movement costume.
SpeedNeeds proves movement correlation before naming anything. The drive-to-reverse rock — foot on the brake, a helper shifting, eyes on the misfire counters — runs on two separate provocations. One observation is not proof.
Correlation proved, the next step isolates the segment: with the engine running, individual harness sections and connectors are manipulated while watching the counters, until the dropout follows one specific segment — reproduced — before it is named. The rock proves correlation; only isolation names the segment.
Stop guessing. Prove the next move. Rock it on command, watch the counters, earn the segment.
Brake-torque provocation is seconds-long, with transmission-temperature caution. This is a diagnostic tap, not a burnout.
Manual transmissions: load the driveline with the clutch against the brake, or rock the vehicle against the parking brake in gear — same principle, no torque converter needed.
Keep hands, tools, and test leads clear — the engine moves suddenly when the gear engages.
If the dropout will not isolate to one segment, it is recorded as movement-correlated with isolation as the explicit next action — not a named part. An unisolated correlation is still a proved fact; it just isn't a repair order yet.
The engine moves, the wires move with it, and a broken connection hides until it flexes. Make it move on command, watch the counters, isolate the segment.
Motion is the test. The counters are the witness. The segment is earned, never assumed.
Stay in the loop
New QTS builds, tools, and garage stories — straight to your inbox. No spam. Unsubscribe anytime.
Read the Privacy Policy.