In RF and microwave wireless systems, accurate power measurement is the most basic requirement. There are a variety of measurement equipment and test methods to choose for power measurement, such as power meter measurement, spectrum measurement and so on. In the actual test work, it shall be ensured that the advantages and limitations of each method will not affect the accuracy of test data.

This paper will discuss the differences between different test methods, one is the use of power meter, the other is the use of spectrometer; Comparison is made from continuous wave (CW), multi tone, modulated signal (32qam) and pulse signal measurements.

A common MXG n5182b is used as the signal source, and its output power is – 20dbm @ 6GHz. The average power meter uses u2000a USB power meter, the peak power meter uses keysight 8990, and the spectrometer uses n9938a handheld spectrometer.

Measure CW signal

The CW signal does not carry any information, so the average power and peak power are the same, and the signal has no bandwidth. Therefore, the measuring equipment can use narrow-band filter to improve the bottom noise of the equipment.

poYBAGFqUAiAN7rAAAG-UJW-hXo862. png

Figure 1: (a) spectrometer (b) power meter (c) channel power measurement option of spectrometer

Fig. 6 is the frequency spectrum of the sweep meter, span = 500KHz. The peak power (also the average power) of the signal is -20.08dbm. Figure 1b shows that the average power measured by USB power meter is -20.00dbm. Figure 7C the average power measured by the channel power measurement option of the spectrometer is -20.09dbm. The channel power measurement option of the spectrometer must set occupied bandwidth for it, that is, set the channel bandwidth of the measured signal. In the measurement in Figure 1C, the occupied bandwidth is set to 100 kHz. However, the continuous wave signal has no bandwidth, and the occupied bandwidth can be set to any value. The difference of the three values is within 0.1dB, and the power meter has the highest accuracy or the lowest measurement uncertainty.

Measuring multitone signals

A multi tone signal with 5 tone and spacing = 500KHz is generated from the signal source, and then connected to the above three test environments respectively. The results are as follows.

pYYBAGFqUA2AWGeFAAGh49Xg5s0407. png

Figure 2: (a) spectrometer (b) power meter (c) channel power measurement option of spectrometer

If the total power of multi tone signal is – 20dbm (10 μ W) , then the power of each tone is -26.98dbm (2 μ W)。 The power meter will measure the total power of all five tones, as shown in Figure 2B, which is -20.13dbm. In Figure 2c, if the occupied bandwidth is set wide and contains all five tones (in this case, the occupied bandwidth should reach at least 2.5MHz), the measured value is 20.2dbm. If the occupied bandwidth is less than spacing, only the power of one tone will be measured, and the measured value is -27.1dbm, which is close to the theoretical value of single tone -26.98dbm. In Figure 2a, the spectrum can measure the power of each tone separately. Theoretically, the power of five tones should be equal, and the measured value is – 27dBm (Figure 1A should be – 27dBm).

If only the power of some tones needs to be measured in the actual test work, because the power meter has no frequency selectivity, only the spectrometer can be used to measure the single tone or specific occupied bandwidth, such as the scene where only the fundamental wave power needs to be measured, or the test scene where im products are tested.

Measure a digital modulation signal (32qam)

Use n5182b to send out a 32qam modulated signal with symbol rate = 5m / s. The average power obtained by the power meter is -20.00dbm (Fig. 3b), and the average power obtained by the channel power measurement option of the spectrometer is -20.07dbm (Fig. 3C).

poYBAGFqUBKAdfWwAAG8TIeEPe0857. png

Figure 3: (a) spectrometer (b) power meter (c) channel power measurement option of spectrometer

As long as the power level and frequency range of the signal are within the capability range of the power meter sensor, the power meter can obtain the total power of the signal with unknown bandwidth. The occupied bandwidth of the channel power measurement option of the spectrometer must be set wide enough to measure accurately. Therefore, the spectrometer can be used to measure the bandwidth of the unknown signal, and then set the occupied bandwidth to the channel power measurement option to test the total power.

Measure a pulse signal

The ratio of the pulse width of the pulse signal (how long the pulse lasts) to the time between the pulses (its period) is called the duty cycle (DF, duty factor). In the case of a given repetitive pulse, the peak power of such waveform is the average power divided by DF.

Fig. 4 is an example of spectrum measurement and average power meter measurement of a pulse signal with a pulse width of 20 μ s. DF is 20%.

pYYBAGFqUBaAYFE8AAGdge6ULEI029. png

Figure 4: (a) spectrometer (b) power meter (c) channel power measurement option of spectrometer

The spectrometer in Figure 4A shows a typical “sin (x) / X” response of the frequency function, and the frequency domain sideband is 1MHz. The average power obtained by the power meter is -27.01dbm. The average power obtained by the channel power measurement option of the spectrometer is – 26.8dbm (Fig. 4C).

Most of the waveforms of the power sideband and the frequency band must be set to 3MHz or above. Increasing the occupied bandwidth above 3MHz will not change the test results. The peak power of this waveform of 20% DF can be calculated by adding 7dB (10log (1 / DF)) to power offset.

How is peak power measured

The block diagram of the average power meter (see Figure 5). The main difference between the peak power meter and the average power meter is that the ADC of the peak power meter has wider bandwidth and higher sampling rate to catch the fast time domain jump in pulses and complex modulation waveforms.

In fact, a typical peak power meter will measure the average power and peak power through two measurement paths.

poYBAGFqUBiAT9T9AABm1k4l1rM505. png

Figure 5: measurement block diagram of simple peak power meter

Because the diode detector is usually broadband, the rapid change envelope of the pulse waveform will be reflected in the output waveform of the diode. The instantaneous power can be obtained by converting and compensating the output level of the diode.

After the filter, the ADC of the peak power meter samples the signal at a rate of up to 1g / S (peak power meter such as keysight 8990). Such a high sampling rate is used to grasp the time-domain shape of the diode output level, from which the signal characteristics such as peak power, pulse width, period, rise time and fall time can be obtained, as shown in Fig. 6.

pYYBAGFqUB2AOT8FAAJhmBXzN9o558. png

Figure 6: pulse signal measured by peak power meter

Summary:

1. The power meter cannot select the frequency component, and can only measure the total power entering the power meter.

2. The sensitivity of the power meter is not high. The minimum power that can be measured by the power meter is basically – 50dbm ~ – 70dBm (depending on the type of probe). The uncertainty is large when measuring small signals, and the measured data needs to be averaged and smoothed for many times.

3. The spectrometer is not a special instrument for measuring power. Compared with the power meter, the power measurement accuracy of the spectrometer is poor. However, when measuring the power of spurious or adjacent channels, the spectrometer shall be selected.

Leave a Reply

Your email address will not be published. Required fields are marked *