O
Member
【Editorial Discussion】
A spectrum-analyser channel-power reading and a power-meter reading on the same modulated signal are often treated as interchangeable absolute-power numbers. They are not, until the integration bandwidth, the detector, and the reference plane are written down. When the two disagree, the usual reflex is to believe the quieter or the more familiar instrument. That hides the term that actually moved.
1) Confirm that both numbers refer to the same plane and the same path loss. An extra adapter, a coupler residual, or a cable that was included on one side only will split the two results without either instrument being “wrong.” Write the plane, and whether mismatch or cal-factor corrections were applied, on both sides before you compare.
2) Channel power is an integral over a declared bandwidth with a declared detector and a declared RBW/VBW (or an equivalent FFT setup). A marker, a peak detector, or a bandwidth that does not match the occupied signal is not the same quantity as a true-average power meter. If the waveform is bursty, also write whether the meter or the analyser was gated, and whether the number is on-slot or a long-term average.
3) Sensor and analyser corrections must match the waveform class, not only a CW cal. A diode or average sensor with a CW cal factor, or an analyser flatness correction that was last checked with a tone, can walk once the signal is wideband. A quick CW cross-check at the same average level is useful, but it does not by itself validate the modulated comparison.
4) Watch the analyser’s front end and the sensor’s range. Compression, an overload flag, or a sensor that is too close to its noise or its peak rating will move one number more than the other. If changing input attenuation or a fixed pad moves the disagreement, you are not yet looking at a DUT power problem.
Open question: the last time channel power and a power meter disagreed on the same modulated signal, what actually closed the gap — reference plane and path loss, RBW / detector / gating, or a cal-factor and range issue? Please share the waveform class and which term you think was dominant; no unverified numbers needed.
A spectrum-analyser channel-power reading and a power-meter reading on the same modulated signal are often treated as interchangeable absolute-power numbers. They are not, until the integration bandwidth, the detector, and the reference plane are written down. When the two disagree, the usual reflex is to believe the quieter or the more familiar instrument. That hides the term that actually moved.
1) Confirm that both numbers refer to the same plane and the same path loss. An extra adapter, a coupler residual, or a cable that was included on one side only will split the two results without either instrument being “wrong.” Write the plane, and whether mismatch or cal-factor corrections were applied, on both sides before you compare.
2) Channel power is an integral over a declared bandwidth with a declared detector and a declared RBW/VBW (or an equivalent FFT setup). A marker, a peak detector, or a bandwidth that does not match the occupied signal is not the same quantity as a true-average power meter. If the waveform is bursty, also write whether the meter or the analyser was gated, and whether the number is on-slot or a long-term average.
3) Sensor and analyser corrections must match the waveform class, not only a CW cal. A diode or average sensor with a CW cal factor, or an analyser flatness correction that was last checked with a tone, can walk once the signal is wideband. A quick CW cross-check at the same average level is useful, but it does not by itself validate the modulated comparison.
4) Watch the analyser’s front end and the sensor’s range. Compression, an overload flag, or a sensor that is too close to its noise or its peak rating will move one number more than the other. If changing input attenuation or a fixed pad moves the disagreement, you are not yet looking at a DUT power problem.
Open question: the last time channel power and a power meter disagreed on the same modulated signal, what actually closed the gap — reference plane and path loss, RBW / detector / gating, or a cal-factor and range issue? Please share the waveform class and which term you think was dominant; no unverified numbers needed.
