Thanks Blazius. With a filter on 3rd gear and pedal 100% i see in that log afr follow lambts (KFLBTS with some multuiplications factors) from 4000rpm.
I hopeed with KFDLBTS =0 and KFFDLBTS =0 to obtain the real KFLBTS value in lambts. BUT....i think is not like this. So the formula is wrong KFLBTS+(KFFDLBTS*KFDLBTS)!
Right?
I will decrease more KFLBTS values and will try to have more fuel from lambts, because i want leave lamfa leaner somewhere at top 12.2.
I will write here the results.

SOME CONCLUSIONs after 1 week with FBSTABGM 0 for axis lower than TABGBTS and 1 when axis >=TABGBTS.
1.KFFDLBTS and KFDLBTS can be 0 - not affect if LAMBTS is active or not , they only multiply. (log 032HS with both maps 0 , log 018h with KFFDLBTS all 1 and KFDLBTS stock)
2. in 018H(BAM) file something activate lambts sooner or later(see logs 018h_1 , 018h_2) ,i dont understood why and what even if TABGBTS is touched from the start of logs (same load , same b_tabgbts , in 018h_1 is lambts is active when is ritch then lamfa). Maybe here someone give an explanation.
3. TLAFA = 0 looks LAMFA si active all the time and BTS never activate! (must make more tests).
4.CWLAMBTS in 032HS is 4 (works with tabgkrm) and in 018H is 0 (work with tabgbts).
5. in 032HS calculated EGT before cat (who is applied for activate lambts) is much bigger than in 018H , aproximatly with 200 degree. I think the calculation of them is different between files. I compared logs EGT with real one gauge read EGT and looks like 018h(BAM) is near real value +- 30degree.
6. in both situation 032hs and 018H, i don't have retards and i confirm and LAMBTS activated, so LAMBTS not depends by KNOCK!

Yes, different calculation
CWLAMBTS.2 you will have a higher egt calculation!

CWLAMBTS.2 = 0 use tabgm
CWLAMBTS.2 = 1 use tabgkrm