What's the difference between spectrometer and oscilloscope?

The differences of analytical performance indexes between oscilloscope and spectrometer are compared from four aspects: real-time bandwidth, dynamic range, sensitivity and power measurement accuracy.

1 real-time bandwidth

For oscilloscopes, bandwidth is usually the measurement frequency range. Spectrum analyzer has the definitions of intermediate frequency bandwidth and resolution bandwidth. Here, we discuss the real-time bandwidth that can analyze signals in real time.

For spectrometer, the bandwidth of the last analog intermediate frequency can usually be used as the real-time bandwidth of its signal analysis. The real-time bandwidth of most spectrum analysis is only a few megahertz, and the wider real-time bandwidth is usually tens of megahertz. Of course, the FSW spectrometer with the widest bandwidth at present can reach 500 MHz. The real-time bandwidth of oscilloscope is an effective analog bandwidth for real-time sampling, usually hundreds of megahertz, up to several gigahertz.

It should be pointed out here that the real-time bandwidth of most oscilloscopes may be inconsistent when the vertical scale is set differently. When the vertical scale is set to be the most sensitive, the real-time bandwidth usually decreases.

In terms of real-time bandwidth, oscilloscope is generally superior to spectrometer, which is especially beneficial for some ultra-wideband signal analysis, especially modulation analysis.

2 Dynamic range

Dynamic range index varies with different definitions. In many cases, the dynamic range is described as the level difference between the maximum signal and the minimum signal measured by the instrument. When the measurement settings change, the instrument's ability to measure large signals and small signals is different. For example, when the attenuation setting of spectrum analyzer is different, the distortion caused by measuring large signals is different. This paper discusses the ability of the instrument to measure large and small signals at the same time, that is, the optimal dynamic range of oscilloscope and spectrometer under appropriate settings without changing any measurement settings.

For spectrometer, the average noise level, second-order distortion and third-order distortion are the most important factors restricting the dynamic range without considering the near-end noise and phase noise. Based on the technical index of mainstream spectrometer, its ideal dynamic range is about 90dB (limited by second-order distortion).

Most oscilloscopes are limited by their AD effective sampling bits and noise floor, and the ideal dynamic range of traditional oscilloscopes is usually less than 50dB. (for R&; S RTO oscilloscope, when 100KHz RBW, its dynamic range can be as high as 86dB).

From the dynamic range, the spectrometer is superior to the oscilloscope. But it should be pointed out here that the same is true for the spectrum analysis of signals. However, the spectrum of the oscilloscope is the same frame of data, and the spectrum of the spectrometer is not the same frame of data in most cases, so the spectrometer may not be able to detect the transient signal. However, the probability of oscilloscope finding transient signal (when the signal meets the dynamic range) is much greater.

3? sensitive

The sensitivity discussed here refers to the level of the minimum signal that can be tested by oscilloscope and spectrometer. This index is closely related to the instrument setting.

For an oscilloscope, when the oscilloscope is set to the most sensitive gear on the Y axis, it is usually 1mV/div. The noise generated by the signal channel of the oscilloscope and the noise generated by the unstable trajectory are the most important factors restricting the sensitivity of the oscilloscope.

As can be seen from Figure 1, due to the increase of sampling points, the spectrum noise floor can be reduced to an ideal level. However, when the signal cannot be reproduced clearly and accurately in the time domain, a large number of clutter will be generated in the frequency domain, which limits our ability to observe small signals.

Figure 1 Sensitivity limit affected by noise

Most oscilloscopes, as shown in figure 1, can stably measure a signal of 0.2mV, corresponding to the frequency domain, which is equivalent to the level of -60dBm. In fact, whether the oscilloscope can accurately measure small signals is not only related to the sensitivity of the vertical system, but also related to the performance of X-axis jitter and trigger sensitivity.

In order to compare the technical indicators analyzed in this paper, the author specially went to R&; S company's open laboratory in Chengdu (thanks to the help provided by Chengdu branch) made an index comparison. Surprisingly, the RTO oscilloscope is excellent in sensitivity index, as shown in the following figure:

Fig. 2 Full-band spectrum of RTO oscilloscope

As can be seen from Figure 2, RTO can accurately measure the signal of -60dBm, and its noise floor is about -80dBm. The most gratifying thing is that there is no large clutter that can affect the sensitivity in the whole frequency band (DC-4GHz), thus greatly improving the measurement sensitivity.

In the absence of clutter, lower noise can be obtained by increasing the number of sampling points. For example, as shown in Figure 3, when the span and RBW are set smaller, the bottom noise of RTO oscilloscope can be reduced to below-100dBm.

Figure 3 Narrow-band spectrum of RTO oscilloscope

From this point of view, RTO can definitely make surveyors change the feeling that "oscilloscope is the chicken rib of frequency domain analysis".

For the spectrometer, the average noise level of the spectrometer can be regarded as the limit of measuring small signals under the conditions of maximum gain and minimum attenuator setting without considering the port mismatch and other factors. Without preamplifier, most spectrum analyzers with good performance can reach-150dBm.

4 power measurement accuracy

For frequency domain analysis, the accuracy of power measurement is a very important technical index. Both oscilloscope and spectrum analyzer have many influences on the accuracy of power measurement. The main impacts are as follows:

For an oscilloscope, the influence of power measurement accuracy includes reflection caused by port mismatch, vertical system error, frequency response, AD quantization error, calibration signal error and so on.

For spectrometer, the influence of power measurement accuracy includes: reflection caused by port mismatch, reference level error, attenuator error, bandwidth conversion error, frequency response, calibration signal error and so on.

We don't analyze and compare the influence quantities one by one here. We compare them by measuring the power of 1GHz frequency signals. Through the measurement comparison between RTO oscilloscope and FSW spectrometer, it can be seen that at 1GHz, the difference between the power measurement values of oscilloscope and spectrometer is only about 0.2dB, which is a very good measurement accuracy index. Because the measuring accuracy of spectrometer at 1GHz is very good.

In addition, the frequency response index of the oscilloscope is also very good in the frequency range, which is less than 0.5dB in the 4GHz range. From this point of view, the oscilloscope is even better than the spectrometer.

Generally speaking, oscilloscope and spectrometer have their own advantages in frequency domain analysis performance, spectrometer is superior in sensitivity and other technical indicators, and oscilloscope is superior in real-time bandwidth. When measuring different types of signals, it can be selected according to the test requirements and different technical characteristics of the instrument.