Recently, I touched two kinds of mixer vector measurement method (R&S and Keysight) and made a comparison with them. All of parameters except reflection response is ok. The responses are as follows.
It is obviously, the reflection response of these two method has huge differences, even in the trend. Although these two kinds of methods have different measurement-step, but the reflection response is important in engineering and this result cannot be accepted.
The process of measurement had been confirmed that there is no problem. Is this problem come from the measurement method itself?