Please hold the body of the calibration standards stationary and only rotate the nuts. Precision connectors can be easily damaged if the pin of the male connector is allowed to rotate in the "pin" of the female connector. That applies to cal standards and cables.
Hello, I would invite to ask Keysight Solution engineer to provide an education update how to use the VNA. You have added a channel, instead of a trace. A channel is a new measurement configuration and have its own calibration set. You could have done correctly your experiment if you would have use {cal}{cal other} and select “Cal All”. Then both channel would have been calibrated and you would not have seen the S22 Problem. For single channel calibration, better to use smartCal and not basic cal.
And also teach your students not to spin RF connectors, especially in cal kits. This means sometimes holding a stubby little end so the center pin does not turn as you tighten the barrel.
Learn about reference planes and your connectors. The Open, or other standard, presents a set of conditions that are well documented and also fed into your VNA (or built in). This allows the analyzer to establish near ideal conditions at the reference plane ... for THAT connector type. All that comes before it, port, test cable, are removed mathematically.
Sorry to say, but you seem to not having understood the very basics of a calibration kit. You can't just take any open/short/load and expect proper results.
Technically you are correct, you should not reuse the standard you have use, because the VNA has computed the error for this particular standard. Better to use another one. But if you don’t have other the you have to know that usually the standard has a delay where the open or short take place. So when you reconnect them you will see the effect of this delay on the smith chart.
Calibration with 'ideal' loads is definitely possible, albeit at the right frequency band where the electrical length of the load/connector is sufficiently small ( typically HF, VHF and lower UHF).
