Why do oscilloscope probes need compensation calibration? What is the underlying reason?-EEWORLD

Collect

Some people may not understand why the probe needs to be compensated.
Let me explain it in detail, hoping that everyone can understand it thoroughly.
All of this starts with understanding the compensation principle of the probe:

Oscilloscope input resistance
The oscilloscope probe cannot send the circuit signal to the oscilloscope. At first glance, it seems that it can be used by directly connecting it.
However, when we use a multimeter to measure the resistance at both ends of the oscilloscope probe, it is nearly 9M ohms, as shown in the figure below:

Multimeter measures the resistance at both ends of the probe X10 gear

. When we look at the oscilloscope, careful friends will find that the 1MΩ input resistance parameter to ground is generally marked next to the BNC input interface of the oscilloscope. Many people may not understand what this represents.

STO1104C oscilloscope BNC input interface

In fact, when using an oscilloscope probe to measure a circuit, because you do not want the connection of the oscilloscope probe to change the working state of the measured circuit itself, the oscilloscope probe must be high-impedance, that is, the input impedance is relatively large (megaohm level). The oscilloscope has a certain voltage input range, but different measurement occasions will have different voltages, so the oscilloscope probe will have different attenuation ratios (1X, 10X, 100X...). Then the simplest way to achieve signal attenuation is to use resistor voltage divider, as shown in the figure below:

In the figure, R1 is the resistor on the oscilloscope probe, and Rin is the input resistor of the oscilloscope. Generally, Rin = 1MΩ, R1 = 99MΩ at 100X, R1 = 9MΩ at 10X, and theoretically it should be 0Ω at 1X, but in fact R1 is about a few hundred ohms, generally within 300 ohms. The

multimeter measures the resistance at both ends of the probe X1 gear

and the input capacitance of the oscilloscope.
Then, according to the resistor voltage divider circuit introduced above, can the oscilloscope be used? No.
As we all know, in reality, any circuit is not an ideal circuit, and there are more or less parasitic parameters. The interface between the oscilloscope and the oscilloscope probe is no exception. Since the oscilloscope interface needs to connect the signal and GND to the oscilloscope probe at the same time (as shown in the figure below, the metal on the outer circle is generally GND, which can play a role of shielding with the outside, and the metal inside is the input signal), a capacitor is formed between the input signal and GND. No matter how the design of the oscilloscope interface is improved, the parasitic parameters of the input capacitance of the oscilloscope cannot be eliminated.
The typical value of the input capacitance of the general oscilloscope is 15pF, 14pF, and 12pF. The figure shows 14pF.

There are R and C, isn't this an RC low-pass filter?

Let's calculate the cutoff frequency of this RC circuit. Considering the 10X gear, R1 = 9MΩ, Rin = 1MΩ, Cin = 14pF, the cutoff frequency is In

this way, all signals higher than 12.64kHz are attenuated to the point where they cannot be seen. So how to solve this problem? Reduce Cin, it is impossible. Physical limitations determine that Cin must exist, and 14pF itself is already a very small capacitance value; reduce R1 and Rin? Too small a resistance will inevitably affect the measured circuit. It seems that there is no other way. But there are always smart people who can find a solution: compensation capacitor. As long as

the compensation capacitor
meets

the requirements, the signal can still be transmitted at a correct 10:1 attenuation ratio at a relatively high frequency without signal distortion.
But a new problem arises. Different oscilloscopes have different Cin. Even if the model is the same, different oscilloscopes may not meet the above proportional relationship with the same probe due to inconsistent manufacturing parameters. Doesn't it mean that a specific probe must be manufactured for each oscilloscope? How to solve the problem of universality? It's very simple, just add a variable capacitor. As shown in the figure Cp. In this

way, you only need to adjust the size of the variable capacitor on the probe and observe the waveform distortion on the oscilloscope to solve this problem.
The oscilloscope generally outputs a 1KHz, 5V (or less) square wave signal, which is used for probe compensation calibration. The signal is often marked with a square wave symbol plus a ground symbol. We can use this signal as a signal source. Start the oscilloscope, as shown in the figure below, connect the BNC of the probe to channel 1, and the other end to the square wave signal output port.

When the following two situations occur, it means that the probe compensation is incorrect, and you need to use the "adjustment stick" to adjust the compensation capacitor on the probe.

Over-compensated waveform

Under-compensated waveform
Use the adjustment stick to turn the screw in the probe screw hole to adjust the compensation capacitor to get the correct waveform.

Some probes also set it at one end of the detection head, as shown in the figure below:

Therefore, when the oscilloscope replaces a new probe or the probe has not been used for a long time, we should calibrate the compensation of the probe.