The CCDF measurement estimates the CCDF of the random variable as defined by the following equation:

where

The following equation defines the CCDF of .

where Pr {e} denotes the probability of the event e.

This equation can be interpreted in logarithmic scale as the probability of the event that instantaneous signal power exceeds the mean power by at least X dB.

The following list describes the basic properties that you can configure for a CCDF measurement:

• RBW: Configure RBW to specify the 3 dB bandwidth over which time domain signal needs to be analyzed. Ensure that RBW does not exceed the instantaneous bandwidth of the signal analyzer. NI recommends that you use no filter or a flat RBW filter. Refer to the Zero Span section of the Spectrum topic for information on the sample rate required by the measurement.
• Measurement Interval and Number of Records: Configure such that the number of samples analyzed by the measurement are enough to get sufficient accuracy in the statistical data reported.
• Threshold: While analyzing signals with dead time, NI recommends to specify a threshold such that only the samples with instantaneous power above this threshold are considered for the measurement.

Probabilities Measurement

Mean power is the linear average of the power of all samples considered for the measurement. When threshold is enabled, only the samples above the threshold are considered. Mean power percentile represents the percentage of samples for which the amplitude is above the mean signal power.

The power, which is relative to the mean power, of 10%, 1%, 0.1%, 0.01%, 0.001%, and 0.0001% of the samples above the mean power is measured. These results help you select the headroom power to be set on the transmitter such that a major portion of the signal is generated with high power without distortion and with only a small percentage of the samples being clipped.

The measurement returns a probabilities trace, which is also called CCDF trace, on which you can interpret any point as follows:

x is the power, in dB, relative to the mean signal power

y is the probability, as a percentage, that any sample in the acquired signal exceeds the mean power by at least x

The CCDF of the power of a complex Gaussian random variable is often used as the reference for comparing the power distribution in the signal under analysis.

The following figure shows the CCDF curve of a QAM 4096 signal.