Full sharing of commercial oscilloscope production skills

Traditional planar or trench MOSFET switches with rise/fall times of 30ns to 60ns are gradually being replaced by power switches such as superjunction MOSFETs with switching times of less than 5ns. To view such fast transitions, an oscilloscope with a bandwidth of at least 1GHz is usually required, but the bandwidth of currently available oscilloscope probes is generally less than 300MHz. In addition, high-frequency voltage and current probes are usually expensive. Therefore, for power engineers in mid-sized companies, the best way is to make their own oscilloscope probes.

To observe fast-changing waveforms, the oscilloscope bandwidth must be at least 1 GHz. Unfortunately, most commercial voltage and current probes cannot operate at such high frequencies.

As modern power supplies operate at higher and higher frequencies, engineers have begun to adopt high-frequency power switch and rectifier technology. Traditional planar or trench MOSFET switches with rise/fall times of 30ns to 60ns are gradually being replaced by power switches with switching times of less than 5ns, such as superjunction MOSFETs, GaN MOSFETs, SiC MOSFETs, and SiC Schottky rectifiers.

To observe such rapid changes, an oscilloscope with a bandwidth of at least 1 GHz is usually required. Unfortunately, most commercial voltage and current probes cannot operate at such high frequencies. The bandwidth of ordinary oscilloscope probes is less than 300 MHz, and the bandwidth of current probes may be only 60 MHz to 100 MHz or even less. In addition, the cost of high-frequency voltage probes is usually more than $12,000, while a slightly better current probe costs at least $4,000. For power engineers working in small and medium-sized companies, there is only one way: make your own probes. Designing and making high-frequency voltage and current probes requires a good understanding of radio frequency, parasitic effects, transmission line theory, and field theory.

Disadvantages of Commercial Probes

Commercial oscilloscope voltage and current probes are extremely robust, ergonomically designed, and very accurate, and they work well in many applications that operate at frequencies well below 1 GHz. The new generation of switching transistors operates at frequencies exceeding 1 GHz, resulting in rise and fall times of less than 5 ns.

The low bandwidth of commercial probes greatly limits measurement accuracy. Engineers are accustomed to slow rise and fall times, so it is easy to ignore missing information. In addition, ordinary probes are connected to the signal source, which will cause distortion. These connections (especially the ground wire) have long sections without shielding. A 4-6 inch (10-15cm) long ground wire may pick up radiated noise from the circuit or other places and inject noise into the coaxial cable to form a common mode signal. This easily ignored signal will be superimposed on the useful signal.

Figure 1 shows a typical commercial voltage probe that contains an unshielded signal or ground lead that forms a loop antenna. The level of noise picked up by this lead is proportional to the loop area as well as the noise energy and noise spectrum. This noise can be observed by simply clipping the ground lead to the probe and then approaching the target circuit board.

Figure 1: A common voltage oscilloscope probe uses a ground lead to clip to the circuit under test.

You can make a 50Ω voltage probe yourself. A 50Ω voltage probe can help you better define and understand what is happening in your circuit. The overall goals of a 50Ω voltage probe are:

• Construct a clean high-frequency signal path from the circuit to the oscilloscope;

• Provide as much shielding as practical along the signal path;

• Ability to control as many parasitic influences as possible.

1:1 shielded coaxial voltage probe

For signals below the maximum rated input voltage of the oscilloscope input, a cut 50Ω BNC coaxial cable can be used as a probe. The unshielded center conductor and shielded tail length should not exceed 1 inch (25cm) to minimize noise pickup. To observe the signal at a specific node, the center conductor can be soldered directly to the node; the ground wire should be soldered to the nearest associated ground, that is, it should not be connected to the ground with a long PCB trace between the probe and the target node. This probe can only provide high-frequency signal shielding from the target circuit to the oscilloscope. The input terminal resistance of the oscilloscope should be 1MΩ. Figure 2 shows the design of this 1:1 shielded probe.

Figure 2: 1:1 shielded coaxial cable-based voltage probe. The inductance (LUS) and ground lead (LG) on the probe will limit the bandwidth, but its small size helps reduce noise pickup.

n:1 50Ω voltage probe

The n:1 probe is mainly used when the signal amplitude (including any spikes) exceeds the maximum rated voltage of the oscilloscope input amplifier. This probe is slightly more complicated to make. Its simplified schematic is shown in Figure 3.

Figure 3: Simplified schematic of an n:1 voltage probe, where the series resistor RS requires some calculation to determine its value.

Therefore, the first and most important step is to determine the size of the sense resistor (RS). This is not as simple as you might think, and there are several factors to consider.

Set the oscilloscope's input terminal resistance to 50Ω, so that the 50Ω terminal resistance inside the oscilloscope becomes the bottom resistance of the voltage divider circuit. You can safely assume that the accuracy of this resistor is better than 0.1%. Its power consumption should not exceed 0.25W. This power rating determines the maximum current value that can enter the oscilloscope input.

Other considerations include:

• Maximum signal amplitude on 50Ω terminal resistor

• Power dissipation in the series sense resistor (RS)

• Load on input circuit

All of these factors must be balanced against each other, and they will determine the gain setting of the oscilloscope input amplifier. If the signal is too low, the oscilloscope input gain must be set to a range less than 100mV. Since the input signal is very close to the noise floor of the input amplifier, the displayed signal will be noisy, resulting in reduced ADC input resolution. The signal may only be captured by the lower four bits of the ADC (assuming an 8-bit ADC), and you will end up seeing quantization steps of the least significant bit (LSB). This situation is difficult to avoid, especially for probes with a high step-down ratio. Figure 4 shows a typical waveform of a 1000:1 50Ω probe.

Figure 4: A low-level oscilloscope trace will often show quantization noise on the input signal.

Figure 5 shows the basic structure of an n:1 voltage probe.

Figure 5: The n:1 50Ω probe has a 1/4W resistor near the probe tip.

The following steps need to be followed when designing an n:1 probe.

First, determine the resistor attenuation ratio based on the desired channel gain setting to achieve a more appropriate oscilloscope signal amplitude (including spikes). Ten times the resistor attenuation ratio is usually selected because the displayed v/div setting only differs in the decimal point position of the input voltage.

The typical input amplitude should not exceed the power rating of the internal input 50Ω termination resistor. To generate the desired channel voltage, current must flow through the 50Ω termination resistor.

The power must be less than the rated power of the terminal resistor:

The calculation formula for the sense resistor (R1) value is:

Now let's examine the power dissipation in the sense resistor.

To load the circuit you want to see, you must understand and determine the impact on the target circuit. If the probe draws too much sense current, it will change the operation of the target circuit, sometimes significantly. The rule of thumb is:

Sometimes the initial considerations are met, but the probe is overloading the target circuit. In this case, you must go back to step 1 and select a sense current that is less than the initial current.