Introduction

In the article "What I Learned While Setting Up the Measurement Environment for Double Pulse Testing," we introduced the basic concepts of double pulse testing and the importance of the measurement environment.

However, once we actually started taking measurements, new questions arose.
"Is this waveform really being measured correctly?"
The display of a waveform does not necessarily mean that the measurement is being performed correctly.
Initially, I myself
"Prepare an evaluation board and observe the waveform." Eon or Eoff It can measure."
That's what I was thinking.

However, in reality,
- Current measurement
Oscilloscope settings
• Probe settings
- Confirmation of measurement conditions
We spent more time setting up the measurement environment than on the measurement itself.

In this article, I will introduce the points I struggled with while actually operating a double-pulse measurement environment, and the following five things I learned from that experience.
1 An oscilloscope was not a "display device." 
2 Measurement was not possible with a current probe alone.
3 I underestimated the importance of the pre-measurement verification process.
4 I underestimated the capabilities of the oscilloscope.
5 There was something I should have checked before watching Eon/Eoff.

An oscilloscope was not a "display device."

Before I started double-pulse measurements,
"An oscilloscope is a device for viewing waveforms."
That was my understanding.
Of course, that's not wrong.
However, in reality,
An oscilloscope was a measuring instrument that directly influenced the measurement results themselves.

At first, the moment the waveform appeared, I thought, "Now I've measured it."
However, in reality,
Input impedance
Trigger settings
• Probe settings
Offset setting
Scale settings
There are many items that need to be checked, such as these.
In particular, after initialization, each channel setting reverts to its default state, so it was necessary to check it each time before taking measurements.
What I learned from this experience is
The fact that a waveform is displayed does not mean that the measurement is being taken correctly.
That's the point.

Figure 1: Main settings checked with an oscilloscope

Even if a waveform is displayed, the input conditions, trigger conditions, and acquisition conditions may not be appropriate. In double-pulse measurements, a reliable waveform can only be obtained when all these settings are properly configured.
<Supplement>
In reality, there were many other items to check besides those shown in Figure 1, such as noise filter settings, GND position adjustment, and self-calibration (Self Cal).
Regarding the GND position, this did not pose a major problem this time because we used an optically isolated probe and a dedicated measurement point provided on the evaluation board.
However, when dealing with high-speed switching waveforms, the way the ground connection is made can affect the superposition of noise and the appearance of the waveform.
Furthermore, self-calibration and noise filter settings may also be affected by the condition of the measuring instrument and the environment.
Therefore, in actual measurements,
The question isn't "Is the waveform displayed?", but "Can we trust that waveform?"
I felt it was important to examine the entire measurement environment from this perspective.

Measurement was not possible with a current probe alone.

When I first started doing double pulse tests,
"If you want to measure electric current, you should use a current probe."
That's what I thought.
However, as we increased the evaluation conditions, we encountered situations where the expected current value exceeded the rating of a typical current probe.
The one used this time TCP0020 While it is a convenient current probe, it has limitations on its rated current.
Forcing the probe to operate under high current conditions can not only prevent accurate measurements but also potentially damage the probe.
Therefore, in this evaluation, CT (Current Transformer Current measurements were performed using a current transformer.

A CT converts the primary current into a secondary signal, and that signal Coaxial cable with BNC connector The signal is input to the oscilloscope via this method.
However, a new problem arises here.
CT is not a genuine current probe that can be automatically recognized by an oscilloscope.
Therefore,
・Current magnification (A/V (Conversion coefficient)
Skew correction value
You will need to manually configure these settings.
In other words,
"CT If you connect it, the current value will be displayed directly."
That's not the case.
So the first thing we did was, TCP0020 and CT This was a comparative measurement using both methods simultaneously.

First TCP0020 Measured under low current conditions that are within the rated range.
- Current value indicated by TCP0020
- Voltage value output by the CT (converted to current value for comparison)
We will compare them.
From the results CT By determining the magnification and setting it on the oscilloscope, you can treat even high-current conditions as correct current values.
Initially, I thought,
"Since the waveform is being displayed, it is being measured."
That's what I thought.
However, in reality,
"The waveform being displayed" and "the current value being correct" are two completely different things.
This was one of the most memorable lessons I learned from this measurement.

Figure 2: Example configuration used for CT magnification confirmation (conceptual diagram)

first TCP0020 Both are measured simultaneously under low current conditions that are within the rating of the (current probe). CT Output and TCP0020 Compare the current values A/V We will calculate the conversion ratio.
<Supplement>
Turns ratio 10:1 The meaning is,
- Toroidal core side: 10 turn
・CT side: 1 turn(1 (penetration)

Figure 3 shows an illustration of the connection between a CT and a coaxial cable with a BNC connector.

Figure 4: Actual photograph of a current probe

Figure 5. Close-up photograph of toroidal core & CT.

I underestimated the importance of the pre-measurement checks.

Before starting the measurement,
"Once you set it up, you can start measuring without any further configuration."
That's what I thought. But in reality, it was different.
Each time a measurement is taken,
- Check the power status
- Checking the VGS waveform
- Check the CT magnification
- Check the skew value
They were doing things like that.

Also, the VGS For measurement and VDS All of the voltage probes used for measurement are optically isolated voltage probes, but their connection methods differ.
VGS measurement side: MMCX connector
・VDS measurement side: 5.08mm Pitch square pin header
We connected to the evaluation board using [this method].
These connection methods cannot be used unless the evaluation board supports them.
The evaluation board used in this study was designed specifically for use with these probes, and its measurement points and connection interfaces were optimized accordingly.

As the measurements progressed,
"Not only the probe, but also the design of the evaluation board significantly impacts measurement quality."
I realized that.

eventually,
The focus is not on the measurement itself, but on reproducing the correct measurement conditions.
I've come to feel that this is more important.

Double pulse test Article titled "What I Learned While Setting Up a Measurement Environment" So, Eon or Eoff We previously discussed the importance of skew adjustment in the evaluation process. Here, we will introduce the skew adjustment fixtures and connection methods that we actually used.

Figure 6: Example of connection between skew adjustment fixture and measuring probe

Figure 7: Illustration illustrating the connection between the skew adjustment fixture and the measuring probe.

For skew adjustment, we used a skew adjustment fixture manufactured in-house.
VGS The measuring probe is MMCX Connect via a conversion adapter, VDS The measuring probe is 5.08mm The connections are made to the spaced square pin connectors. Before measurement, these adjustments and checks were performed, followed by waveform evaluation.
<Glossary>
Fixture: A specialized jig (device) used to fix and connect the object to be measured or the probe.
Skew: Time difference that occurs between multiple measurement signals
VDS: MOSFET Drain-source voltage
VGS: MOSFET Gate-source voltage
Pulse Generator: An abbreviation for pulse generator. A measuring instrument that generates a reference pulse signal.
MMCX: A standard for small, high-frequency coaxial connectors.
Pitch: Distance between the centers of adjacent terminals

Figure 8: Comparison of waveforms before and after skew adjustment

Before skew adjustment, there was a time difference between the IDS, VGS, and VDS waveforms. By correcting the time axis of each waveform through skew adjustment, Eon and Eoff can be evaluated more accurately.

Figure 9 VGS/VDS measurement points on the evaluation board

The connection point of the voltage probe varies depending on the object being measured and the device package.
The evaluation board used in this experiment had dedicated connection points, but we checked the connection destinations and probe configuration each time before taking measurements. This kind of preparation was essential for obtaining reliable waveforms.

I underestimated the capabilities of oscilloscopes.

Through this evaluation, ADC I also had the opportunity to learn about resolution.
With an oscilloscope,
8-bit mode
・12-bit mode
You can select this option.
Initially, I thought that the default settings would be sufficient.
However, when you actually compare the waveforms, 8bit and 12-bit Then there was a difference in how the waveforms looked.

In 8-bit 256 step, 12-bit So 4096 Voltage can be expressed in stages.
Therefore 12-bit This allows for more detailed observation of minute waveform changes and noise components.
On the other hand, if you look only at the resolution 12-bit While this is advantageous, in reality, it is also necessary to consider the balance between sampling conditions and acquisition time.

In this measurement, we wanted to prioritize higher waveform quality, so ultimately 12-bit The settings have been adopted.
Before I started taking double-pulse measurements, ADC I didn't pay much attention to resolution and often used it with the default settings.
However, through this experience,
Oscilloscope settings don't just affect the appearance; they affect the information that can be obtained.
I realized that.

Table 1 Differences between sampling rate and resolution in oscilloscopes

With the Tektronix MSO64 oscilloscope used in this test, measurements were possible even with the default 8-bit setting.
However, after actually comparing them, I felt that the 12-bit setting made it easier to perceive waveform changes and noise components, so I ultimately adopted the 12-bit setting.

There was something I should have checked before watching Eon/Eoff.

When I first started learning about double pulse testing, what interested me most was, Eon, Eoff, and reverse recovery properties (Qrr)
However, in reality, Eon and Eoff These are not things you see first. The final value That's it.

In order to perform accurate measurements,
Oscilloscope settings
- Current measurement
・CT magnification setting
Skew correction
・VGS check
We need to check each one individually.
And only when those things are done correctly can Eon and Eoff It will come to have meaning.

The difficulty in double-pulse measurement lies not in the object being measured itself, but in correctly constructing the measurement environment.
This time,
"To what extent can we trust the waveforms obtained from that measurement environment?"
I learned that this is important.

Through this series of tasks, I realized that Eon/Eoff Rather than "measuring",
"Create an environment where Eon/Eoff can be measured correctly."
I felt that this was more important.

The construction of the measurement environment and various settings described so far will ultimately Eon or Eoff This is preparatory work for evaluating it. By actually performing the measurements, VGS, VDS, IDS The switching behavior of the device can be observed from the waveform.

Figure 10: Representative waveform obtained after environment setup.

These are representative measurement waveforms obtained after setting up the measurement environment, configuring the CT magnification, performing skew correction, and carrying out various verification procedures.
In double pulse testing, VGS, VDS, and IDS can be observed simultaneously to evaluate the switching behavior of the device and the effects of differences in circuit configuration.

lastly

When I first started learning about double pulse testing, I focused on Eon and Eoff themselves. However, I learned that in reality, with double pulse testing, it's more important to build a reliable measurement system before even looking at the measurement results.
I hope this article will be helpful to those who are planning to implement double pulse testing.

For related articles, please see the link below.
​Five things I wish I had known before starting double-pulse measurements: Lessons learned while setting up the measurement environment

Inquiry

If you are interested in the content introduced in this article, please feel free to contact us.