Power sensor or spectrum analyzer: To measure a radio system and network, in addition to the measuring device used, the corresponding measuring method also plays an important role.

For increasingly interconnected plants and equipment, efficient radio systems and networks must be developed. For such systems to work well, the RF and microwave power must be measured. This includes the selection of measuring instruments and the corresponding measuring methods.

In our article, we show the consequences of various average power measurements using the example of continuous wave, multi-frequency, 32-QAM (quadrature amplitude modulation), and pulsed signals. We compare the measurement method with the help of a power sensor and the power measurement via a spectrum analyzer.

Using the example of the vector signal generator MXG from Keysight, a test signal with a frequency of 6 GHz is generated. The level is -20 dBm. The power sensor used was a USB power sensor U2000A connected to a FieldFox N9938A Spectrum Analyzer with option 302. The spectrum analysis measurement method used the same FieldFox configured with an internal channel power meter (CPM, Channel Power Meter, option 310). The system was connected to the signal generator via a coaxial cable.

## Continuous-wave signals or CW signals

Continuous-wave signals or so-called CW signals = Continuous Waves are not modulated. Thus, their average power is equal to the peak power. The message consists of a single frequency and has very low bandwidth. Therefore, you can use a narrowband filter to increase the sensitivity and thus reduce the inherent noise of the meter.

The exact value of the measurement bandwidth is irrelevant to a continuous wave signal as long as the signal bandwidth is much smaller than the measurement bandwidth. However, a low measurement bandwidth extends the measurement time, whereas a larger measurement bandwidth increases the noise and thus reduces the measurement accuracy.

## How to measure a multi-frequency signal?

In the example of a *multi-frequency* signal, the output power of the signal generator is distributed to five frequencies at intervals of 500 kHz. If the total output power is -20 dBm (10 μW), every single rate has a strength of -26.98 dBm (2 μW). If a sensor measures the average ability, specifically, in this case, the total power, this is -20.13 dBm for all five frequencies. The CPM shows a measurement result of -20.2 dBm, provided its measurement bandwidth is set to capture all five individual rates.

In the present case, at least 2.5 MHz. If the CPM measurement is repeated with a measurement bandwidth of only 200 kHz, only the middle tone at 6 GHz is recorded, and reading of -27.1 dBm is obtained, which is close to the theoretical value of -26.98 dBm for a single frequency. If you want to measure the power of single frequencies or measure a specific bandwidth, you can do that only with a spectrum analyzer, because power sensors are not frequency selective.

The peak power of the multi-frequency signal differs from the average power. This is due to the phase relationship between the individual frequencies. To measure the peak power, you need a unique sensor that can measure peak power and a filter that is wide enough to handle all affected rates simultaneously.

## Measurement of a digitally modulated signal

A power sensor provides a correct average power signal for a signal of unknown bandwidth, as long as the sensor’s measurement range covers the power of the signal and its frequency range. With a CPM in a FieldFox, on the other hand, you have to set the correct measurement bandwidth for accurate results.

For this, one uses first the normal function of the spectrum analyzer, with which one finds out the bandwidth of the unknown signal. After determining the bandwidth, switch to CPM, and then set this measurement function to the correct bandwidth.

In the case of pulsed signals, the ratio of pulse width = time in which the signal is switched on at the time between two consecutive pulses = the period as duty factor (Duty Factor). The peak power of such a signal is the average power divided by the duty cycle. This assumes that the signal is periodic.

An even wider setting did not bring any further change in the reading. The peak power of this 20 percent duty cycle signal can be calculated by adding the average power of 6.99 dB (= log _{10} (1 / duty cycle)). If you measure -27 dBm for the average power and add 6.99 dB to it, you get -20 dBm. That’s the power the signal generator was set to.

## Leave a Reply