You can also measure loss using a single-port analyser.
The underpinning maths is here:
On 06/08/2016 07:03, Richard (Rick) Karlquist wrote:
No, this isn't the real deal. A 2 port vector network analyzer
with full 2 port calibration is the real deal. The wattmeter
approach has many error effects that are difficult to account
for. The fact that the meters read identical when cascaded
does not mean the subsequent measurement is perfect.
The dummy load may have a mismatch. The characteristic
impedance of the line is not exactly 50 ohms and is actually
complex in general. Etc.
TowerTalk mailing list