“The main purpose of conducting double-pulse testing is to obtain the switching characteristics of power semiconductors, which can be said to accompany the entire life cycle of power devices from R&D and manufacturing to application. Many things can be done based on the device switching waveform obtained from the double-pulse test, including: verifying the device design scheme and proposing improvement directions through the analysis of the switching process, extracting the switching characteristic parameters to make the device specification, calculating the switching loss and reverse recovery loss as the power supply Thermal design provides data support, comparison of switching characteristics of devices from different manufacturers, etc.
“
The main purpose of conducting double-pulse testing is to obtain the switching characteristics of power semiconductors, which can be said to accompany the entire life cycle of power devices from R&D and manufacturing to application. Many things can be done based on the device switching waveform obtained from the double-pulse test, including: verifying the device design scheme and proposing improvement directions through the analysis of the switching process, extracting the switching characteristic parameters to make the device specification, calculating the switching loss and reverse recovery loss as the power supply Thermal design provides data support, comparison of switching characteristics of devices from different manufacturers, etc.
Measuring the effect of delay
The measured signal will experience two delays during the measurement process, and the difference in delay experienced by different signals will have a certain impact on the measurement results. The primary delay is the delay of the analog front end of the oscilloscope. The difference between the delays between different channels of the oscilloscope is at the ps level, and the switching process of the power device at the ns level and the us level can be ignored. The other time is the delay of the probe. The direct delay difference of different probes is in the ns level. At this time, it has a significant impact on devices with fast switching speeds, especially for SiC and GaN, which have been gradually popularized in recent years. The device impact is even greater.
We illustrate the effect of the measurement delay by taking the measurement of the switching process of a SiC MOSFET as an example.
In the figure below, the blue waveform is the waveform obtained before delay calibration is performed on the probe (before calibration), and the red waveform is the waveform obtained after delay calibration is performed on the probe (post-calibration waveform).
According to the theory, during the turn-on process, when IDS starts to rise, a voltage drop will be generated on the parasitic inductance of the loop, which will make VDS drop. The rise of IDS and the drop of VDS should start at almost the same time. In the pre-calibration waveform, when IDS starts to rise, VDS remains unchanged and starts to fall after a 5.5ns delay. This situation is obviously inconsistent with the theory, and it can be inferred that the IDS signal leads the VDS signal at this time. After calibration, this problem was solved.
Theoretically, during turn-off, the peak of VDS should be near the zero-crossing of IDS. In the pre-calibration waveform, VDS reaches the highest value only after IDS zero-crosses 5.5ns. After calibration, this problem was also solved.
At the same time, we also sorted out the switching characteristic parameters, including: turn-on delay td(on), turn-on time tr, turn-on energy Eon, turn-off delay td(off), turn-off time tf, and turn-off energy Eoff. It can be seen that the turn-on delay td(on) and turn-off delay td(off) have a difference of about 2ns before and after calibration, the turn-on time tr and turn-off time tf are almost unchanged before and after calibration, and the turn-on energy Eon before calibration is 395.31uJ compared with 147.53 after calibration The uJ is 1.67 times larger, and the turn-off energy Eoff of 20.28uJ before calibration is 71.3% smaller than 70.54uJ after calibration.
It can be seen that the delay has a very obvious impact on the analysis of the switching characteristics of the device and the calculation of parameters. Before measuring, we can use the following four methods to perform delay calibration.
Method 1: Simultaneously measure the square wave signal of the oscilloscope
Calibrating the delay difference between different voltage probes is very simple, just measure the same voltage signal at the same time, and then calibrate according to the measurement result.
We can use the square wave signal that comes with the oscilloscope, usually on the side Panel or front panel of the oscilloscope. It can be seen that although the two probes are measuring the oscilloscope’s own square wave, the waveform displayed on the oscilloscope screen is delayed by 5.6ns in the 1-channel measurement channel ahead of the 2-channel measurement channel. Then we can set the delay of channel 1 to -5.6ns in the channel setting menu of the oscilloscope, or set the delay of channel 2 to +5.6ns to complete the calibration.
Method 2: The oscilloscope’s own function calibration
For a fixed type of voltage or current probe, the delay is basically fixed, and the delay calibration can be performed directly as long as the probe model is known. Now the functions of the oscilloscope are becoming more and more powerful, and the model of the probe can be identified through the interface of the probe, so that the oscilloscope can directly and automatically calibrate the device delay value.
Method 3: Time-lapse calibration fixture
Following the idea of measuring the same signal at the same time to calibrate the delay difference of different voltage probes in Method 1, and measuring the voltage and current signals that change at the same time, the delay difference between the voltage probe and the current probe can be calibrated. Many oscilloscope manufacturers have launched voltage and current probe delay calibration fixtures based on this idea, which is convenient for engineers to use directly without having to make their own calibration sources. Such as Tektronix’s deskew pulse generator signal source TEK-DPG and power measurement offset correction fixture 067-1686-03.
Method 4: Utilize Device Switching Characteristics
If limited by the equipment at hand, we can also calibrate by the characteristics of the device switching process mentioned above. According to the theory, the rise of IDS and the fall of VDS should start at almost the same time during the turn-on process. Then we can carry out double-pulse test under different test conditions, read the delay between IDS and VDS during each turn-on process, and then calibrate after averaging.
About Tektronix
Headquartered in Beaverton, Oregon, USA, Tektronix provides innovative, precise and easy-to-use test, measurement and monitoring solutions that solve problems, unlock insight and drive innovation. For more than 70 years, Tektronix has been at the forefront of the digital age. Welcome to join our innovation journey, please log on: tek.com.cn
The Links: CM200TXPA-24T DP10D1200101804