Glass fibre insulated probe during the double pulse test

Wide bandgap (WBG) semiconductor materials such as silicon carbide (SiC) and gallium nitride (GaN) are not only characterised by high temperature and voltage resistance, but also exhibit properties such as low losses and fast switching frequencies. However, in order to fully utilise the potential of these advanced materials, precise test and measurement techniques are of crucial importance. Especially in double-pulse tests, optical insulation probes not only ensure the safety of the test process, but also improve the accuracy and reliability of the measurements. This article explains why optical isolation probes are essential for double-pulse testing.

What is a double pulse test?

The double pulse test (DPT) is an experimental method for evaluating the switching performance of power electronic devices such as IGBTs (Insulated Gate Bipolar Transistors) or MOSFETs (Metal-Oxide-Semiconductor Field-Effect Transistors). In this test, two voltage pulses of short duration are applied to the device, simulating its switching behaviour in actual circuits. It is used to measure and analyse the switching characteristics of the device, to optimise the device drive and application design, as well as for fault diagnosis and validation of simulation models.

The choice of oscilloscope and probes in DPT

In the example of the half-bridge gate driver circuit for MOSFETs, we need to test Vds, Id and Vgs of the lower MOSFET and also observe Vgs of the upper MOSFET. The choice of the Micsig MHO high-resolution oscilloscope series with 500 MHz bandwidth, 3GSa/s sampling, ≤1 % accuracy and all 4 channels supporting simultaneous observation of the upper and lower MOSFET switches perfectly fulfils the test requirements for DPT (double pulse testing).

To accurately measure the Id waveform, ensure that the current probe used has sufficient bandwidth. Use the high frequency AC/DC current probe CP series from Micsig, which offers a bandwidth of up to 100 MHz, an accuracy of 1 % and a resolution of up to 1 mA and supports measurements up to 30 A. For test requirements with larger currents, you can use the Rogowski AC probe RCP series. However, many users may have the following question: "We have the Micsig DP series high voltage differential probes that performed well when testing silicon devices, even measuring voltages up to 7000 V with a bandwidth of 500 MHz. If we now switch to GaN and SiC devices, the probes should theoretically fulfil the bandwidth requirements of these devices. You can test the lower MOSFET, but why do you always encounter problems when testing the voltage of the upper MOSFET?

By analysing and comparing the data from the figure above, we found that both SiC and GaN switches achieved switching speeds in the nanosecond range. This significant advantage reduces the energy consumption of switching power supplies, but also poses a major challenge for testing. In a half-bridge circuit, the Vgs of the high-side MOSFET fluctuates between switching the Vds voltage on and off, which can change from zero volts to thousands of volts in just a few nanoseconds. The combination of high voltage and high frequency leads to a significant increase in the higher order harmonic components. The differential voltage Vgs of our device under test is often only a few tens of volts and there is significant common mode interference from the higher order harmonic components of Vds. When measuring, we need to suppress this common mode interference as much as possible. This requires the test set to maintain a high common mode rejection ratio (CMRR) even in the high frequency range. Using the DP series as an example, the CMRR is >-70 dB at 100 kHz, >-40 dB at 20 MHz and >-26 dB at 120 MHz. For differential probes, this CMRR is already considered excellent among peers. However, it is far from sufficient to fulfil our requirements for measuring the gate-source voltage (Vgs) of the high-side MOSFET at high frequencies. We need a device that maintains a very high CMRR even in the high frequency range.

Comparison of high voltage differential probe and optically isolated probe

In terms of the effects of CMRR on testing, let's make a comparison to look at the problems caused by high voltage differential probes during testing and the comparison of high CMRR probes.

Test method: The device under test is a SiC switch with high and low gates operated with a Vce voltage of approx. 500 V. Simultaneously use a high voltage differential probe (Micsig DP) and an optically isolated probe (Micsig SigOFIT) connected to the high side Vge signal for double pulse tests.

The figure above shows the test result. The white signal represents the result of the high voltage differential probe. It can be observed that strong oscillations occur during the rising edge of Vge, making it almost impossible to recognise the original waveform. We previously used a high-voltage differential probe to test the upper gate signal (Vge) of a device with a Vce voltage of up to 800 V, and the oscillations exceeded the turn-off voltage of the SiC, which significantly affected the engineer's evaluation. Whereas the red waveform in the figure was obtained with an optically isolated probe, resulting in significantly less signal interference. In independent tests with an optically isolated probe, there is almost no interference. The interference observed here is caused by the high voltage differential probe affecting the optically isolated probe. In fact, an optically isolated probe has a lower baseline noise compared to the high voltage differential probe, providing greater accuracy and the ability to measure larger common mode voltages. How is this achieved?

Advantages of the optically isolated probe

Micsig's exclusive SigOFIT™ technology selects the appropriate attenuator for the signal magnitude under test before testing, allowing the differential mode signal from ±0.01 V to ±6250 V to be tested at full scale. While adapting to a wide range of tests, it improves test accuracy (up to 1%), reduces the noise floor and improves the signal-to-noise ratio.

Modellvergleich	MOIP100P	MOIP200P	MOIP350P	MOIP500P	MOIP800P	MOIP1000P
Bandbreite	100 MHz	200 MHz	350 MHz	500 MHz	800 MHz	1GHz
Anstiegszeit	≤3,5 ns	≤1,75 ns	≤1ns	≤700 PS	≤438ps	≤350 PS
SMA-Eingangsimpedanz	1MΩ; \|\| 10pF	1MΩ; \|\| 10pF	1MΩ; \|\| 10pF	1MΩ; \|\| 10pF	1MΩ; \|\| 10pF	1MΩ; \|\| 10pF
Ausgangsspannung	±1,25V	±1,25V	±0,5V	±0,5V	±0,5V	±0,5V
Differenzspannungsbereich	1X: ±1,25 V 2X: ±2,5V 10X: ±12,5V 20X: ±25V 50X: ±62,5V 500X: ±625 V 1000X: ±1250V 2000X: ±2500V 5000X: ±6250V	1X: ±1,25 V 2X: ±2,5V 10X: ±12,5V 20X: ±25V 50X: ±62,5V 500X: ±625 V 1000X: ±1250V 2000X: ±2500V 5000X: ±6250V	1X: ±1,25 V 2X: ±2,5V 10X: ±12,5V 20X: ±25V 50X: ±62,5V 500X: ±625 V 1000X: ±1250V 2000X: ±2500V 5000X: ±6250V	1X: ±0,5V 2X: ±1V 10X: ±5V 20X: ±10V 50X: ±25V 500X: ±250 V 1000X: ±500V 2000X: ±1000V 5000X: ±2500V 1000X: ±5000V	1X: ±0,5V 2X: ±1V 10X: ±5V 20X: ±10V 50X: ±25V 500X: ±250 V 1000X: ±500V 2000X: ±1000V 5000X: ±2500V 1000X: ±5000V	1X: ±0,5V 2X: ±1V 10X: ±5V 20X: ±10V 50X: ±25V 500X: ±250 V 1000X: ±500V 2000X: ±1000V 5000X: ±2500V 1000X: ±5000V
Rauschen	<450μVrms	<450μVrms	<450μVrms	<450μVrms	<450μVrms	<450μVrms
Ausbreitungsverzögerung	15,42 ns (2 m Kabellänge)	15,42 ns (2 m Kabellänge)	15,42 ns (2 m Kabellänge)	15,42 ns (2 m Kabellänge)	15,42 ns (2 m Kabellänge)	15,42 ns (2 m Kabellänge)
Stromversorgung	USB Typ-C, DC: 5 V	USB Typ-C, DC: 5 V	USB Typ-C, DC: 5 V	USB Typ-C, DC: 5 V	USB Typ-C, DC: 5 V	USB Typ-C, DC: 5 V
DC-Verstärkungsgenauigkeit	1 %	1 %	1 %	1 %	1 %	1 %
Gleichtaktspannungsbereich	60kVpk	60kVpk	60kVpk	60kVpk	60kVpk	60kVpk
Kabellänge	2 M (Std.) (anpassbar)	2 M (Std.) (anpassbar)	2 M (Std.) (anpassbar)	2 M (Std.) (anpassbar)	2 M (Std.) (anpassbar)	2 M (Std.) (anpassbar)

At a bandwidth of up to 1 GHz, the noise of the SigOFIT probe can be within 0.45 mVrms. Even in the frequency range of 1 GHz, the CMRR remains above 100 dB. Therefore, when using an optically isolated probe to measure the high-side Vgs, the effects of common mode noise need not be considered, perfectly solving the problem of insufficient CMRR in high-voltage differential probes.

In addition, due to the length of the leads (generally around 20 cm), differential probes can act like antennas and pick up external magnetic field interference. Given the extremely high switching speed of GaN, the magnetic fields it generates can trigger oscillations when passing through the input of high voltage differential probes, sometimes exceeding certain limits and causing instant burnout or explosion of the GaN devices. Optically isolated probes, on the other hand, use MCX or MMCX connections with very short leads, which virtually eliminates the antenna effect. The parasitic capacitance is in the range of a few pF, effectively eliminating safety risks from parasitic effects during the test.

Conclusion

To summarise, the optical isolation probe has indeed outperformed the differential probe in all aspects of performance. For users who need to perform double pulse tests, the optically isolated Micsig probe is the best choice.