After talking with Dave B. and Patrick T. about the set point values I gave them for CP3,4 auto overfills, I discovered that CP3 stops overfilling when thermocouples see < -30C. CP4 is set to -60C. I cannot reason why I would have set them differently, but now I know -30C is too high, and I'm concerned that last Friday's overfill never actually overfilled. Attached is a plot of Sunday night's manual fill right at the pump, where I stood for 50 minutes and verified LN2 coming out of exhaust (the final dip you see in plot). For most of those 50 minutes, the exhaust was just cold gas reading below -50C.
Also attached are the fill plots from last Wed. and Fri. You can see the usual shallow slope in Wed. plot followed by the LN2 finale. Friday's fill fell immediately to something less than -30C which could have just been cold gas.
We will change the set point to -100C for both CP3,4.
I think your observation in the data of how long the thermocouples spent below -30C (the auto fill "stop" set point) on Sunday night's manual fill in the absence of LN2 actually exiting the exhaust makes a convincing argument for abbreviated fill(s) on Friday (and Wednesday?).