doubled injections possible on negative fueling transients

crank to run transition


Note that the duration of these two intervals is approximately the same:


What's happening is that the first small injection is scheduled, then the injection phase is updated, then at the next tooth its scheduled AGAIN. Since we target the end-of-injection timing, the first injection got scheduled expecting the duration of the longer cranking pulse, not the shorter running pulse.

Start by reproducing the crank-to-run transition described in the issue and compare the two injection intervals shown in the screenshots. Trace where the first injection is scheduled and where the injection phase is updated; done means the transition no longer schedules the same negative fueling transient twice.

Written by the indexing model from the issue text.