From radar to 5G communications networks, common use-cases that require measuring the RF power of a system include proof-of-design, satisfying regulatory specifications, adhering to safety limits to protect against the dangers of high-power RF radiation, system efficiency, and component protection.

A diode-based power sensor uses high-frequency diodes to rectify the RF signal to a DC voltage signal. Proportional to the amplitude of the RF signal, this voltage is captured near the signal source. Instruments like an RF power meter can measure and scale the DC voltage to reveal desired power measurements. The relation between the DC voltage and power result depends on whether the diode operates in the square-law, transition, or linear region.

Peak power sensors have a small smoothing capacitance and use a low-impedance load across the smoothing capacitors to discharge rapidly when the RF amplitude drops. This means that peak power sensors can deliver quick rise times and wide video bandwidths.

Peak power sensors have triggering capabilities, which allow them to synchronized pulse measurements. When choosing a peak power test solution, engineers should consider a sensor's video bandwidth to accurately track a signal's envelope power, rise time to capture a pulse's rising edge, time resolution to verify pulse shape and timing, and crest factor/statistical measurements to assess component linearity.

Average power sensors operate in the square-law region where the DC voltage output is closely proportional to the square of the applied RF voltage. These sensors, therefore, deliver accurate and reliable average power measurements despite the presence of modulation.

Video bandwidth and the ability to track a signal's profile are less of a concern when selecting an average power sensor, since these instruments are only used to measure average power over time and instantaneous envelope results are not required.

Envelope power is a signal’s amplitude variations from modulation or distortion as a function of time averaged over one or a few cycles of the RF carrier signal.

Peak envelope power is the maximum point or peak amplitude of a signal’s envelope power.

Average power defines the average power level over the pulse repetition interval, which includes the signal burst and the interval before the next pulse.

Pulse average power is the average power level of the signal burst over the pulse width.

Video bandwidth describes the frequency range that a sensor can track a signal’s fluctuations in envelope power.

A sensor with limited video bandwidth will not always be able to track envelope power fluctuations. With a response speed that is too slow, various measurement errors will occur since there will be points in time when the carrier power is unknown. For example, a sensor with insufficient video bandwidth can affect envelope power, peak envelope power, and average power measurements.

Rise time is the time interval for a signal’s leading edge to change or rise from a certain low value (e.g., 10% of the pulse magnitude) to a certain high value (e.g., 90% of the pulse magnitude).

Sensors with sufficient rise times will be able to provide clean and crisp measurements of leading edges of high-speed signals, such those used in radar and 5G time division duplex (TDD) applications, for accurate characterization and analysis.

Video bandwidth and rise time are inversely proportional. An increase in video bandwidth decreases the rise time, meaning the system can respond to faster signal changes (decreased rise time) over a wider frequency range (increased video bandwidth).

The formula for a calculating a sensor’s rise time and video bandwidth can be defined by:

Rise Time (ns) = 0.35 / Video Bandwidth (GHz) or Video Bandwidth (GHz) = 0.35 / Rise Time (ns)

For example, a sensor with 195 MHz VBW would be able to track a signal with a rise time of about 2 ns (Rise Time = 0.35 / 0.195 GHz ≈ 1.79 ns). When specified, the sensor’s performance would be more conservatively defined as having a rise time of less than 3 ns. Using the same example, a sensor with a rise time of about 2 ns (but specified as less than 3 ns) equates to a video bandwidth of 195 MHz (Video Bandwidth = 0.35 / 1.79 ns ≈ 0.195 GHz).

A pulsed RF signal periodically changes between “on” and “off” states. The RF carrier signal is at a non-zero power level during the “on” state, while the “off” state reduces the signal strength completely in ideal conditions.

The pulse repetition frequency is the number of pulses from a repeating signal that are transmitted per second. It’s typically measured in pulses per second and captured in Hz.

The pulse repetition interval is the time interval between two successive pulses measured at the same point on both pulses.

Pulse width is the time interval between a single pulse’s leading and trailing edges. It measures how long a single pulse is in its “on” state.

The pulse period is the time interval between two successive pulses measured at the same point on both pulses.

Duty cycle is the ratio of a pulse’s “on” state to the total period of the signal. It is calculated by taking the ratio of the pulse width over the pulse repetition interval.

The pulse envelope is a trace of the change in amplitude or shape of each pulse, which is often displayed on power measurement instrumentation. An ideal pulse envelope is completely rectangular with vertical leading and trailing edges. Real-world pulsed RF signals do not behave ideally and experience distortions that can impact performance.

Pulses that are “on” for long periods of time are more susceptible to pulse distortions such as droop, which ultimately affect system performance. For example, next-generation radar systems transmit long pulse width signals. Range will be reduced and target acquistion time affected if the highest level of waveform fidelity isn’t maintained.

Pulse droop is a signal distortion from an ideal, flat pulse top where the amplitude reduces between the beginning and end of the pulse. Pulse droop can be caused by factors like thermal effects.

Overshoot is a signal distortion where the waveform surpasses the desired value immediately after the pulse leading edge. This contrasts from ideal behavior where a pulse’s zero rise time seamlessly transitions to its top amplitude, creating a flat-topped rectangular shape.

Crest factor, also referred to as the peak-to-average power ratio (PAPR), is a figure of merit used to better understand the shape of a waveform. It takes the ratio of the peak amplitude to the average amplitude or root-mean-square (RMS) value.

The crest factor of a pulsed signal can be calculated by taking the peak envelope power over either the average power during the pulse repetition interval (average power) or the average power of the pulse (pulse average power).

A unitless calculation, the formula for crest factor can be expressed as:
Crest factor = V_peak/V_avg
Where:
V_peak = Peak voltage
V_avg = Average voltage
Often, crest factor is converted to decibels (dB) with the following equation:
Crest factor = 20log(V_peak/V_avg).

Signals with a high crest factor are more dynamic and have severe peaks that rise far above average power levels. Compare this behavior to signals with low crest factor, which have a lower peak amplitude relative to the waveform’s average power.

High crest factor waveforms are often seen in applications (e.g., 5G networks) that use higher order modulation schemes like orthogonal frequency-division multiplexing (OFDM). Adding together various independently modulated subcarriers causes extreme power peaks in multi-carrier OFDM systems, while destructive interference between subcarriers, among other factors, can keep average power levels low in comparison.

Signals with high crest factors can be affected by an amplifier when it is operated in its non-linear region. Non-linear amplifiers clip the highest peaks of modulated waveforms, causing signal distortion, symbol errors, and bit errors. Increasing the input back-off value (IBO) can improve linearity by reducing the amplifier’s input power, but this process also reduces amplifier efficiency and range.

The complementary cumulative distribution function (CCDF) is a statistical representation of how frequently a specific crest factor occurs, showing the percentage of time or probability (y-axis) that a signal's crest factor is equal to or greater than a certain crest factor value (x-axis).

Let’s say 0.01% is the probability percentage of interest for a CCDF plot. If the point where 0.01% on the y-axis intersects with the CCDF curve is equal to 10 dB (x-axis), then the signal’s crest factor is equal to or greater than 10 dB for 0.01% of the time. Another way to interpret this result is that the signal’s crest factor is equal to or less than 10 dB for 99.99% of time.

Amplifier linearity is verified if the CCDF input and output curves follow a near identical path with similar crest factor values. Observing a notable reduction in the output signal’s crest factor, however, reveals waveform clipping, system nonlinearities, and amplifier compression.