The heating pad came already glued to my aluminum bed, but the magnetic foil isn’t glued, it still has the sticker covering the glue before applying.
The path you could go is interesting, but I would question the same points: most people can achieve incredible prints with klipper’s compensation, why can’t we?
My opinion is something is giving innacurate probe reads. If I try to do probe_calibrate I see the value for z_compensation, and hence Z, change multiple times in the same place.
At one point, I changed /usr/share/klipper/klippy/extras/probe.py and removed the math for z_compensation - this actually gave me a better result on the first layer (still bad, not terrible though).
But… it begs the question: why am I even doing this?