L
Admin
[Technical article / FAQ, rfforum Editorial Team]
Average-power sensors disagree once the signal is pulsed, bursty, or crest-factor rich. Teams then argue about "which sensor is right" instead of which quantity (average, peak, heating) each technology actually reports.
1. What each family measures
- Thermistor / thermoelectric: closer to true heating / average; slow; great for CW cal transfer, poor for short pulses without a clear average definition
- Diode sensors: fast, need modulation calibration / crest-factor awareness; video bandwidth and diode linearity set when "average" mode is still honest
- Calorimeter / reference standards: top of hierarchy for CW; transferring that number to a diode on a pulsed waveform is a method problem, not a single cal factor
2. Freeze list before you sign a power number
- Waveform: CW, pulsed duty, OFDM crest factor — named on the same line as the watt reading
- Sensor model, head, and whether the meter used a modulation cal / pulse mode or CW cal only
- Mismatch / cal factor vs frequency, and whether the pad or coupler in front was in the cal chain
- For PA efficiency work: are you comparing DC input to the same definition of RF out the gateway uses?
3. Habit
Never archive a power number without waveform class + sensor type + cal mode. "We used a good Keysight/R&S head" is not a metrology sentence on a pulsed PA sheet.
Discussion: on your last pulsed or OFDM power disagreement, did the fix land in sensor type, modulation cal, or a crest-factor definition mismatch — and do you require waveform class on every power traveler?
— rfforum 编辑部 / Editorial Team
Average-power sensors disagree once the signal is pulsed, bursty, or crest-factor rich. Teams then argue about "which sensor is right" instead of which quantity (average, peak, heating) each technology actually reports.
1. What each family measures
- Thermistor / thermoelectric: closer to true heating / average; slow; great for CW cal transfer, poor for short pulses without a clear average definition
- Diode sensors: fast, need modulation calibration / crest-factor awareness; video bandwidth and diode linearity set when "average" mode is still honest
- Calorimeter / reference standards: top of hierarchy for CW; transferring that number to a diode on a pulsed waveform is a method problem, not a single cal factor
2. Freeze list before you sign a power number
- Waveform: CW, pulsed duty, OFDM crest factor — named on the same line as the watt reading
- Sensor model, head, and whether the meter used a modulation cal / pulse mode or CW cal only
- Mismatch / cal factor vs frequency, and whether the pad or coupler in front was in the cal chain
- For PA efficiency work: are you comparing DC input to the same definition of RF out the gateway uses?
3. Habit
Never archive a power number without waveform class + sensor type + cal mode. "We used a good Keysight/R&S head" is not a metrology sentence on a pulsed PA sheet.
Discussion: on your last pulsed or OFDM power disagreement, did the fix land in sensor type, modulation cal, or a crest-factor definition mismatch — and do you require waveform class on every power traveler?
— rfforum 编辑部 / Editorial Team
