ECM unusual behaviour

1227727 ECM $8D
Hello everyone,
I've been tuning the car for a new engine and after uploading a new bin file (I was looking to get rid of low throttle jerking/choking) adjusting AE),the PE would not activate until a MAP of 88Kpa. I didn't even change anything to PE with the new bin file. I've even burned the old bin where it was ok and the same behaviour. The settings are 192F high temp threshold for PE, then above 2000rpm the activation is 30%TPS for cold and 14% for hot. It activates only at 40-55%TPS exactly at 90kPa according to the Tuner Pro datalog. Below 2000rpm the threshold is 45% so the ECM is not using that everywhere.
The XDF I use is S_AUJP v1026 and already burned a lot of bins with no problems.

The car was in a shop where they replaced the radiator and the coolant temperature is now 160F minimum and 185F while driving; 195 when idling. Before that It would run 185 through 210F. So what changed was it now runs cooler or perhaps the mechanics did something in the shop. Another potential cause of some malfunction could be disconnecting and reconnecting the ECM. Because of the American Style battery terminals it causes intermittent connection while I turn the bolt. I have mounted a capacitor that keeps the power for 2 seconds but that can be also a source of trouble as it could keep the voltage longer at a level where it is not good. For example 5V.
Another thing I've noticed is the AFR is dropping when I let off the gas as PE activates there. That could be simply mean fuel is still pumped a bit after I close the throttle valve, so I guess it's normal. No pops at the exhaust.
The car is a C4 corvette 1991 TPI.
This runs on the 1227727 ECM $8D AXCN

I plan to do another test then put the old ECM back in to see if it does the same. I also have 2 MEMCALS that I use alternatively and swapping them didn't solve anything.
I've attached a log if that helps
Thank you

PS
I've ran the car with even older bins on the other ECM and I get the same behaviour. However, I've noticed there are many older log files where it does the same thing. I don't actually know where PE activates from the log files but I can see the drop of the target AFR. That can be due to AE as well. I guess I need to study the fuel algorithm for some more time. Based on what I know so far I could just set the VE higher by a percentage plus the expected BLM compensation in the needed areas, or just figure out how to activate PE where I want.