L
Admin
"What is the power?" sounds like a one-button question. Between sensor physics, mismatch and cable loss, absolute power is one of the easiest RF measurements to get systematically wrong by 0.3 to 1 dB.
1. Sensor families
- Thermal (thermocouple / thermistor): averages true heating; excellent for CW and complex waveforms if the head is rated; slow; limited dynamic range; expensive.
- Diode (peak / average capable heads): wide dynamic range, fast; calibration curves assume a waveform class. Crest factor and modulation can bias "average" modes if you are outside the calibrated envelope.
If you change from CW cal to a high-PAPR OFDM burst without checking the head mode, trust drops.
2. Mismatch factor (the quiet error)
Delivered power P_del = P_inc * (1 - |Gamma_load|^2) under simple models, but the meter reads something closer to incident power at its reference plane depending on head design and cal factor definition. Always read the head manual for whether cal factors are referenced to a matched load.
Rule: if DUT |S11| is poor, a good source match and a good head match still leave residual uncertainty. A tuner or pad in front of the sensor improves effective match at the cost of loss you must account for.
3. Traceability chain
Power meter reading
→ sensor cal factor at frequency (table or EEPROM)
→ adapter / cable loss (measure, do not guess)
→ reference plane move (fixture de-embed if needed)
→ DUT mismatch correction if your procedure requires delivered power
Skipping cable characterization is the most common 0.2 to 0.5 dB lie in "quick" benches.
4. Practical checklist
- Zero / cal the head at the temperature you will use
- Use the cal factor for the actual frequency, not the nearest pretty number
- Prefer a 3 to 10 dB pad on the sensor for better match when measuring amplifiers into a notching load
- For modulated signals, confirm average vs peak mode and video bandwidth settings
- Cross-check once against a second method (SA channel power with known RBW corrections, or a second head)
5. When the SA is better
Relative measurements, spectral shape, ACLR. Absolute channel power on an SA needs its own amplitude accuracy story (attn, preamp, flatness). Do not treat SA dBm as gospel without that chain.
Absolute power is a metrology problem wearing an RF jacket. Write the reference plane on the traveler.
1. Sensor families
- Thermal (thermocouple / thermistor): averages true heating; excellent for CW and complex waveforms if the head is rated; slow; limited dynamic range; expensive.
- Diode (peak / average capable heads): wide dynamic range, fast; calibration curves assume a waveform class. Crest factor and modulation can bias "average" modes if you are outside the calibrated envelope.
If you change from CW cal to a high-PAPR OFDM burst without checking the head mode, trust drops.
2. Mismatch factor (the quiet error)
Delivered power P_del = P_inc * (1 - |Gamma_load|^2) under simple models, but the meter reads something closer to incident power at its reference plane depending on head design and cal factor definition. Always read the head manual for whether cal factors are referenced to a matched load.
Rule: if DUT |S11| is poor, a good source match and a good head match still leave residual uncertainty. A tuner or pad in front of the sensor improves effective match at the cost of loss you must account for.
3. Traceability chain
Power meter reading
→ sensor cal factor at frequency (table or EEPROM)
→ adapter / cable loss (measure, do not guess)
→ reference plane move (fixture de-embed if needed)
→ DUT mismatch correction if your procedure requires delivered power
Skipping cable characterization is the most common 0.2 to 0.5 dB lie in "quick" benches.
4. Practical checklist
- Zero / cal the head at the temperature you will use
- Use the cal factor for the actual frequency, not the nearest pretty number
- Prefer a 3 to 10 dB pad on the sensor for better match when measuring amplifiers into a notching load
- For modulated signals, confirm average vs peak mode and video bandwidth settings
- Cross-check once against a second method (SA channel power with known RBW corrections, or a second head)
5. When the SA is better
Relative measurements, spectral shape, ACLR. Absolute channel power on an SA needs its own amplitude accuracy story (attn, preamp, flatness). Do not treat SA dBm as gospel without that chain.
Absolute power is a metrology problem wearing an RF jacket. Write the reference plane on the traveler.