Dang! - My measurments were not full characterization, but in-circuit with low I_c. So, not peak and I_c might explain most of the variation!
Now that I've got the gear, I could do proper measurements if I can find some older ones.
Thanks for spotting that out!