开云体育

As there is some discussion on what is limiting the S21 measurement I did some tests.
First is a single sweep of S21 without a DUT , no averaging, without applying the isolation calibration.
Then the same measurement but with exponential averaging.
This shows a very coherent phase in the higher frequencies and the S21 amplitude increases steadily with the frequency
There is a 10dB jump down at 300MHz as the measurement switches to harmonics mode as this is not removed by calibration as the leakage follows a different signal path.

Then I measured S21 of a 50dB attenuator with exponential averaging.
The amplitude of the first 500MHz is fairly flat at -50dB, then the amplitude goes down till 650MHz and then up and from around 750MHz the amplitude follows the same levels as in the isolation measurement till 1300MHz where it deviates again.
The dip at 650MHz is not caused by the attenuator (I verified using a 3GHz VNA) but by the combining of the leakage signal and the signal through the 50dB attenuator as the signals are anbout 180degree out of phase.

my NanoVNA has S21 dynamic range at about 70 dB. Below 300 MHz it is usable down to 90 dB with some noise.

From 0 dB to 50 dB it has almost straight line with attenuators. 60 dB has a little deviations. 70 dB has more deviation. 80 dB a little noisy.

The problem with CH1 port is that it has not so good VSWR. It drops down to 42 ohms at 450 MHz and to 40 ohm at 900 MHz.