Yury Kuchura EU1KY
Just look at the curves for magnitude ratios (Ua/Ub) versus Z measured with EU1KY and NanoVNA bridges in 1...1000 Ohm range (see the picture attached).
As you can see, NanoVNA produces almost flat line at high Z. The difference is 0.9 dB between 500 and 1000 Ohms. I don't say about 1000 and 2000 Ohms difference, which is hard to measure and makes a very noisy R/X graph for high-Z loads. At the same time, there is a steep peak around 50 Ohms. This makes measurements around 50 Ohms very unstable because of too low signal level. You should know that the mixer cannot fully suppress common mode voltage present on its both inputs. Let's say, ordinary 612 mixer does ~20..25 dB only without special efforts, and whatever you do for balancing it, you can do it on a single frequency only, and only for a short period of time because of drift. This yields a poor precision in 40..63 Ohms range where Ua/Ub is above 20 dB, and common mode voltage makes a significant error in magnitude and phase measurements.
EU1KY bridge provides 4.65 dB difference between 500 and 1000 Ohms, and the line is relatively slightly curved due to high enough resistor in series with DUT (after many experiments 200 Ohms was chosen as a most optimal one that provides good enough impedance measurement precision in the 0..2000 Ohm range). At the same time there is no common mode voltage problem, at all. Yes, the input impedance is not 50 Ohms, but this is not significant compared to achieving the main goal: more or less precise measurement of low and high impedances.