A brief discussion on the classification of oscilloscope probes and their impact on measurement

For an oscilloscope, its input interface is generally a BNC or 3.5mm coaxial interface. If the output of the device under test uses a similar coaxial interface connector, it can be directly connected to the oscilloscope via a cable; if the signal to be tested is on the PCB board, or the signal under test does not use a coaxial connector, a corresponding oscilloscope probe is required.

Anyone who has used an oscilloscope will have come into contact with a probe. Usually, the oscilloscope we are talking about is used to measure voltage signals (there are also those that measure light or current, which are first converted into voltage measurements through corresponding sensors). The main function of the probe is to lead the measured voltage signal from the measurement point to the oscilloscope for measurement. The following figure shows various oscilloscope probes.

Most people pay more attention to the use of the oscilloscope itself, but ignore the choice of probe. In fact, the probe is the intermediate link between the measured signal and the oscilloscope. If the signal is distorted at the probe, then no matter how good the oscilloscope is, it will be useless. The figure below is an example. The rise time of a common 500MHz passive probe is about 700ps. When using this probe to test a signal with a rise time of 530ps, even if the influence of the oscilloscope bandwidth is not considered, the rise time of the signal after passing through the probe has become 860ps. Therefore, the influence of the probe on the measurement cannot be ignored.

For oscilloscopes and probes with Gaussian frequency response, the bandwidth of the measurement system consisting of the probe and oscilloscope can usually be calculated using the following formula:

For a flat-response oscilloscope and probe, the bandwidth of the measurement system it forms depends on the part with the smallest bandwidth. It can be seen that the probe and its connection method have a great impact on the test system.

In fact, the design of a probe is much more difficult than that of an oscilloscope, because the inside of an oscilloscope can be well shielded and does not need to be frequently disassembled, while the probe must not only meet the requirements of detection convenience, but also ensure at least the same bandwidth as the oscilloscope, which is much more difficult. Looking back at the development history of oscilloscopes, many high-bandwidth real-time oscilloscopes did not have probes with corresponding bandwidths when they first appeared, and probes with corresponding bandwidths were usually launched some time later.

To choose a suitable probe, the first thing to do is to understand the impact of the probe on the test, which includes two parts: the impact of the probe on the circuit being tested and the signal distortion caused by the probe itself. The ideal probe should have no impact on the circuit being tested and no distortion on the signal. Unfortunately, no real probe can meet these two conditions at the same time, and usually some compromises need to be made between these two parameters.

In order to consider the impact of the probe on the measurement, we can usually simply equate the input circuit of the probe to the R, L, C model shown in the figure below (the actual model is much more complex than this). During testing, we need to analyze this model together with our circuit under test.

First, the probe itself has input resistance. Just like the principle of measuring voltage with a multimeter, in order to minimize the impact on the circuit being measured, the input resistance Rprobe of the probe itself is required to be as large as possible. However, since Rprobe cannot be infinite, it will produce a voltage divider with the circuit being measured, causing the actual measured voltage to be different from the real voltage of the probe. This situation is often encountered in the testing of some power supplies or amplifier circuits. In order to avoid the influence of the probe resistance load, the input resistance of the probe is generally required to be at least 10 times greater than the source impedance and load impedance. The input impedance of most probes is between tens of k ohms and tens of M ohms.

Secondly, the probe itself has input capacitance. This capacitance is not deliberately added, but is the parasitic capacitance of the probe. This parasitic capacitance is also the most important factor affecting the bandwidth of the probe, because this capacitance will attenuate the high-frequency component and slow down the rising edge of the signal. Usually, the parasitic capacitance of high-bandwidth probes is relatively small. Ideally, the parasitic capacitance Cprobe of the probe should be 0, but it is not possible in reality. Generally, the input capacitance of a passive probe is between 10pf and several hundred pf, and the input capacitance of an active probe with a higher bandwidth is generally between 0.2pf and several pf. Due to the existence of parasitic capacitance, the input impedance of the probe (note, not the DC input resistance) will decrease with frequency, thereby affecting the bandwidth of the probe.

The following figure shows the curve of the input impedance of two commonly used probes changing with frequency. The input impedance of both probes is high impedance under DC conditions: the most commonly used high-impedance passive probe with a bandwidth of 500MHz can have an input impedance of 10MΩ under DC conditions, and the input impedance of another single-ended active probe with a bandwidth of 2GHz is 1MΩ under DC conditions. However, since the high-impedance passive probe on the left has a larger parasitic capacitance, its input impedance decreases faster with increasing frequency. When the frequency reaches 70MHz, its input impedance is already much smaller than that of the active probe with a smaller parasitic capacitance. Therefore, the input parasitic capacitance has a great impact on the probe bandwidth.

Secondly, the signal input by the probe will also be affected by parasitic inductance. The input resistance and capacitance of the probe are relatively easy to understand, but the inductance at the input end of the probe is often overlooked, especially when measuring at high frequencies. Where does the inductance come from? We know that where there is a wire, there will be inductance. There must be a wire connecting the probe and the circuit under test, and the signal return must pass through the probe's ground wire. The ground wire commonly used in oscilloscope probes usually has an inductance of about 1nH for a 1mm probe length. The longer the signal and ground wires, the greater the inductance value. As shown in the figure below, the parasitic inductance and parasitic capacitance of the probe form a resonant circuit. When the inductance value is too large, the resonant frequency is very low, and it is easy to generate high-frequency resonance under the excitation of the input signal, causing signal distortion. Therefore, the length of the signal and ground wires needs to be strictly controlled during high-frequency testing, otherwise ringing is likely to occur.

Before understanding the structure of the probe, we also need to understand the structure of the oscilloscope input interface, because this is where the probe is connected. The oscilloscope's input interface circuit and the probe together constitute our detection system.

Most oscilloscope input interfaces use BNC or BNC compatible interfaces (some high-bandwidth oscilloscopes use some specially designed interfaces, such as 2.92mm or 1.85mm coaxial interfaces). As shown in the figure below, many common oscilloscopes have 1M ohm or 50 ohm switchable matching resistors at the input. There are many types of oscilloscope probes, but the matching of oscilloscopes only has two options: 1M ohm or 50 ohm. Different types of probes require different matching resistor forms.

From the perspective of voltage measurement, in order to minimize the impact on the circuit under test, the oscilloscope can use a high input impedance of 1M ohm, but the bandwidth of the high impedance circuit is very sensitive to the influence of parasitic capacitance. Therefore, the input impedance of 1M ohm is widely used for measurements below 500M bandwidth. For higher frequency measurements, 50 ohm transmission lines are usually used, so the 50 ohm matching of the oscilloscope is mainly used for high-frequency measurements. Traditionally, most oscilloscopes with a bandwidth below 100MHz on the market only have 1M ohm input because they are not used for high-frequency measurements; most oscilloscopes with a bandwidth of 100MHz to several GHz have a switching option between 1M ohm and 50 ohm, taking into account both high and low frequency measurements; oscilloscopes with a bandwidth of several GHz or higher are mainly used for high-frequency measurements, so most of them only have 50 ohm input.

In a broad sense, test cables are also a type of probe, such as BNC or SMA cables, and this type of probe is both cheap and high-performance (provided that the quality of the cable is not too poor). However, when using test cables to connect, there must also be a BNC or SMA interface on the circuit being tested, so the application occasions are limited, mainly used for RF and microwave signal testing. For digital or general signal testing, special probes are often still required. The figure below is a classification of some commonly used probes in oscilloscopes.

Oscilloscope probes can be divided into passive probes and active probes according to whether they need power supply, and can be divided into voltage probes, current probes, optical probes, etc. according to the type of signal measured. The so-called passive probe means that the entire probe is composed of passive components, including resistors, capacitors, cables, etc.; while active probes generally have amplifiers inside, and the amplifiers need power supply, so they are called active probes.