Guide updates and tuning notes by email → Follow us
Plan a swap

Home › Q&A › Tuning

Why does AutoTap show less timing advance than commanded without any knock retard?

Your knock learn factor is sitting below full. The ECM saw knock earlier and is now blending spark from the High Octane table toward the Low Octane table in those cells — that learned pull stays in place and keeps subtracting timing even when the knock retard PID reads zero. Log Knock Learn Factor to confirm.

Last updated

Who this applies to

  • Computers: P01, P59, E38, E67
  • Engines: LS1, LS6, LM7, LQ4, LQ9, LS2, LS3, LS7
  • Not for: Gen V LT (different knock-learn strategy)

Gen III P01/P59 (1997-2007)

  1. Log Spark Advance, Knock Retard, and Knock Learn Factor in VCM Scanner over the cells in question.
  2. Commanded spark = High Octane cell minus (High minus Low) times the learn factor, per cell — not the whole table. Diff 10, factor 0.75 removes 2.5 degrees
  3. The Maximum Knock Retard table caps how far the blend can pull timing in those cells. Factor 0.5, Max Retard 4 caps pull at 4 degrees not 5
  4. A reflash resets the learn factor to zero so the ECM starts back on the High table. Resets to 0 on flash

Gen IV E38/E67 (2005-2013)

  1. Watch the Knock Learn Factor PID in the scanner under Engine > Timing Retard; on these controllers it runs 0-1, low octane at 0, high at 1. 0-1 scale
  2. Expect it to take real driving to swing back up to the high table after knock stops. 30-50 miles to swing up from low octane map

Careful

  • Some knock activity on 87 octane with a stock tune is normal GM logic; do not chase it to zero by gutting the High table.
  • If you cannot hear knock but KR persists, suspect false knock before pulling more timing.

More in tuning fundamentals, or build a checklist for your swap with the tune checklist tool.

Go deeper

Check this first

  1. Log Knock Learn Factor alongside Knock Retard and Spark Advance; a factor below 1 in the problem cells with zero KR is the smoking gun.
  2. Check the High vs Low octane tables actually differ in those cells; if they are identical the blend removes nothing.
  3. Verify fuel octane matches what the High table was tuned for; trace knock on 87 octane is normal stock behavior.

Try this, then this

  1. Fill with the octane the High table was built for and drive; the learn factor recovers on its own with no knock. — Removes the cause so the factor climbs back toward the High table. Worked if: Knock Learn Factor returns to full and commanded spark matches the High table.
  2. Reflash the calibration to reset the learn factor, then re-test on good fuel. — Flashing clears the learned blend instantly instead of waiting out the recovery drive. Worked if: Timing immediately matches the High table until/unless knock returns.
  3. Pull timing out of the High Octane table in the knock cells so the learn factor stops having to. — A table that never knocks never engages the learn blend. Worked if: No KR and learn factor stays at full across the drive.
  4. Hunt false knock: loose header bolts, exhaust contact, noisy valvetrain near the knock sensors. — The sensors cannot tell real knock from mechanical noise. Worked if: KR disappears at the same timing values that used to trigger it.

Key numbers

WhatValue
Knock learn factor blend mathDiff of 10 between tables, factor 0.75 = 2.5 degrees removed
Flash resets learn factorResets to 0
Maximum Knock Retard capFactor 0.5 with Max Retard 4 caps reduction at 4 degrees
E38 knock learn factor scale0-1, 0 = low octane map, 1 = high octane map
Learn recovery distance30-50 miles

Time and cost: No parts; 30-50 miles of mixed driving for the learn factor to recover, or one reflash.

Something wrong on this page? Tell us