Loosen chi2 cuts in KPi Asymmetry selection for 2023

Histograms of early '23 showed a quite poor selection efficiency of charm mesons for the detection asymmetry lines. This is in part due to the (very very) tight fdchi2 and IPCHI2 cuts that cut especially hard when the resolution is poorer.

I have tuned the rates to be about or below 1kHz inclusive for the (500kHz) HLT1-filtered nobias MC sample. For 2024, we can tune the IPCHI2 cuts once more. Note that the bandwidth should be alright as (1) it's not persistreco, and (2) a lot of the rate is inclusive, not exclusive.