| Reading your post makes my head hurt. So instead of an answer for each question I'll generalize how the ECM operates. '7747 & '8746:
When the ECM goes into PE mode the commanded AFR is changed to the value from the PE AFR vs RPM table. As the RPM changes the ECM will track the RPM AFR value of that table. There is no locking of the AFR at the entry to PE mode.
In PE mode: Learn is turned off and the INT is locked at 128. If the BLM is < 128 it is ignored (128 is used). If the BLM is >= 128 it is then used. The upper BLM RPM limit is only used during learn. IOW, no learning takes place above that limit (such as above 3600 RPM).
However, the use of the BLM table extends past this boundry. So when in PE mode (or not in PE mode), there is no upper RPM limit for the use of the BLM cells. They extend to infinity. This is also true for the upper and lower MAP limits. These limits only affect learn, not the usage during fueling corrections. And, the BLM cell is use will also change to reflect the changing MAP/RPM values.
The best way to tune the upper VE table areas including high MAP, is to force open loop, set the BLM limits to 128, and use a WB. Change the VE table to match the commanded AFR. For the commanded AFR you will need to flatten the open loop AFR tables (two of them). This way you will know what the AFR is being commanded to be (unless using a Lockers high speed data acquisition system, then you will have the commanded AFR along with the WB AFR at the same time).
Do the same with the PE mode AFR table, make it the same value at all RPM's. Then you will know what the commanded AFR is. And be able to change the VE table to have it match up to the WB. No need to lock out PE mode, that is not a good idea.
RBob. | 
