As described here, I've build an RF Powermeter.
After aquiring an HF signal generator, a used HP 8647A, I started to do some testing regarding the precision of the RF powermeter.
Comparision HP 8647A vs. R&S URV5 with URV5-Z5
First of all, I did a comparison between the HP 8647A and my R&S URV5 with the URV5-Z5 power sensor. The power sensor is specified from 1nW to 20mW with a frequency range from 10MHz to 18GHz.
1nw=-60dBm and 20mW=+13dBm - so the range of my powermeter is within the range of the URV5-Z5 power sensor. The resulting diagram is shown below:
Here you can see, that the power sensor and the signal generator are nearly perfectly in sync between 144MHz and 1GHz in a power range from 0dB down to -50dBm.
Calibration and accuracy of my powermeter
Here are the calibration constants for my powermeter which I've programmed in my meter:
const PROGMEM float kFactors[] = {
2.80342581E+01, -3.40466467E-02, 5.36059447E-06, -9.64826574E-10, // 50 MHZ
2.81602896E+01, -3.43993780E-02, 5.36926411E-06, -9.35643168E-10, // 145 MHZ
2.73776126E+01, -3.35128770E-02, 4.76795974E-06, -8.13901646E-10, // 433 MHz
2.63845320E+01, -3.25466666E-02, 4.19419955E-06, -7.11984824E-10, // 1000 MHz
2.63845320E+01, -3.25466666E-02, 4.19419955E-06, -7.11984824E-10, // 2000 MHz
2.63845320E+01, -3.25466666E-02, 4.19419955E-06, -7.11984824E-10, // 3000 MHz
};