Why are the ripple noise measurements of different oscilloscopes always different?

For the same Power Supply, there are always differences in using different oscilloscopes to measure ripple and noise values. Even using different probes can affect the measurement results. What is the reason?

First, the difference between ripple and noise

Ripple

Since the switching tube of the switching power supply operates in a high-frequency switching state, each switching process, the electric energy is "pumped" from the input end to form a charging and discharging process on the output capacitor, thereby causing fluctuations in the output voltage. Moreover, the frequency of this fluctuation is the same as the switching frequency of the switch tube. This fluctuation is the output ripple, which is an AC component superimposed on the output DC. The amplitude of the ripple is the peak-to-peak value between the peak and the valley of the AC component.

noise

The noise is caused by the switching power supply itself, which is caused by the sharp pulse generated when the switch is turned on and off. The frequency of the noise is much higher than the switching frequency. The size of the noise voltage is largely related to the switching power supply. The topology, the winding of the Transformer, the parasitic parameters in the circuit, the external electromagnetic environment during testing, and the wiring design of the PCB.

Based on this, the distinction between ripple and noise is well understood. Let's start by discussing the measurements.

Second, the reasons for the difference between different oscilloscope test results

If the differences between different oscilloscopes are large, there are generally several reasons:

1, not using 20M bandwidth limit

The noise values introduced by different bandwidths are different, and the noise value directly affects the Pk-Pk value measurement.

2, no grounding spring

The long ground wire of the probe and the loop formed by the probe at the contact of the circuit board are too large, and it is easy to introduce a large amount of electromagnetic interference in the space into the measuring circuit.

3, probe frequency compensation is not corrected

The probe itself is in a state of good compensation, and the amplitude measurement itself is not accurate.

have to be aware of is:

It must not be calibrated by which brand or model measurement result. It is considered that a certain brand is right, a certain model is correct, and an oscilloscope is very expensive, so it is right. These measurement mentalities are not desirable. It is quite normal for the test results to be within a certain range. After all, the oscilloscope has only an 8-bit ADC, and the vertical resolution is poor. In addition, the amplitude-frequency response curves of different oscilloscopes are slightly different.

Third, a little reflection in the measurement process

At this time, people can't help asking: Why do you have to limit the bandwidth of 20M? Is there any scientific basis?

This 20M is a simple hardware filter circuit designed by early oscilloscopes to avoid excessive power supply output due to high frequency interference. If you want to go deeper, personally think that 20M can only be an experience value, not a theoretical value.

Fourth, measure ripple noise in a more accurate way

At present, the new digital oscilloscope function is getting stronger and stronger. We don't have to limit our thinking to the 20M frame limit, and we can directly get more accurate signals through digital filtering. Generally need to use the three functions of the oscilloscope: FFT, digital filter, statistical measurement

(1) FFT operation function.

Through FFT operation, the frequency domain characteristics of the waveform can be effectively analyzed, and the power, RMS, phase and other characteristics of each frequency component existing in the waveform can be visually observed, and the harmonic components and distortions in the system under test can be effectively applied. , noise characteristics in the power signal, etc. ZDS2000 series oscilloscopes are equipped with 4Mpts FFT. The frequency resolution of the 1GSa/s FFT equivalent sampling rate can still be as fine as 250Hz, which can accurately and quickly analyze the source of signal interference.

(2) Digital filter function.

It uses a IIR or FIR digital filter to digitally filter the signal under test. With less computational effort, it is possible to have good flatness in the passband, sufficient attenuation in the stopband, and sufficiently small stopband ripple. The ZIR2024 oscilloscope's standard IIR digital filtering function can select high-pass or low-pass filtering, and the cutoff frequency can be selected from 100Hz to 100MHz. As shown in Figure 1 below, the ZDS2024 oscilloscope sets the amplitude-frequency response curve of the low-pass Butterworth filter with a normalized cutoff frequency of 0.08:

Figure 1 The amplitude-frequency response curve of the low-pass filter

(3) Automatic measurement function.

Not only can you measure the waveform of the channel source, but you can also measure the waveform after mathematical or digital filtering. The “true sense” parameter measurement statistics standard on the ZDS2000 series oscilloscopes will measure all the waveforms captured on the screen to obtain the current value, maximum value, minimum value and average value, standard deviation, and number of measurements. By observing the statistical maximum and minimum values, the user can quickly understand the possible abnormalities in the waveform, and can quickly evaluate the signal characteristics by observing the average value and the standard deviation.

2, the typical use of each function

FFT function: can effectively analyze the frequency domain range of noise in the signal;

Digital filtering function: can effectively filter the noise components in the signal;

Automatic measurement function: It can be used to measure various parameter indicators of waveforms after mathematical operation or digital filtering.