My network analyzer manual defines tracking error as:

"the vector sum of all test setup variations in which magnitude and phase change as a function of frequency"

A response calibration normalizes all measurements with respect to a reference measurement. Can you explain how the "division" operation with respect to reference trace measurement has the operative effect of decreasing magnitude and phase change as a function of frequency?

Can you give me numerical example at a single frequency of how a response cal can correct for tracking error - magnitude and phase?

With respect to a one port calibration:

How can the tracking error terms alone (in isolation), e01 and e10, be directly measured and calculated? If the calculation is associated with a group delay term, please show the relation between group delay and e01 and e10.

"the vector sum of all test setup variations in which magnitude and phase change as a function of frequency"

A response calibration normalizes all measurements with respect to a reference measurement. Can you explain how the "division" operation with respect to reference trace measurement has the operative effect of decreasing magnitude and phase change as a function of frequency?

Can you give me numerical example at a single frequency of how a response cal can correct for tracking error - magnitude and phase?

With respect to a one port calibration:

How can the tracking error terms alone (in isolation), e01 and e10, be directly measured and calculated? If the calculation is associated with a group delay term, please show the relation between group delay and e01 and e10.

In S21, if you measure uncorrected a thru connection, which is know to have zero loss, the S21 should be 0 dB (linear 1)

In the measurment, if you really measure -3 dB (linear 0.7) due to cable loss. To compute tracking you do the math:

Tracking = S21meas/S21actual (always the math is done in linear, thus 0.7/1 = 0.7

And to do the error correction you do the math:

S21actual = S21meas/Tracking (0.7/0.7=1)

For S11, it is a bit more difficult and the tracking effect can be confused by the source match term. But this is all well describe in many VNA user manuals.

In general, you cannot separate e01 and e10; if you use a power meter, you can determine the magnitude of each (we do this in scalar mixer cal (smc) in the PNA). But currently it is not possible to measure the phase of one path independently in a VNA.