Originally Posted by nota4re

All good questions, and you are right that I don't know all of the details behind how it captures. You bring up an interesting point - maybe it can only detect misfire from lack of spark?? However, with enough misfires to stumble the car, I'd think that a lean condition should be detectable from the O2's. Perhaps the ECU tolerance level is not exceeded, however, and the stumble gets back on track before the ECU gets angry. This is the tolerance you are speaking of when removing the intake tube. No fault code may be generated but if you were to try to drive like that you wouldn't get very far before a fault code. (I've noticed this when we forgot to re-connect the MAF sensor connector.)

The most convincing evidence of fuel starvation is the fact that the problem doesn't occur if kept above 3/4 full. I am simply trying to broaden the thinking to see if it could be something else - hopefully a simplier/easier fix.

Originally Posted by CO_VaporGT_09

Same if I disconnect the intake tube while it's running, and the MAF/ECU metering tries to correct, the car stumbles, or even dies, but I don't see any codes or check engine lights.