How to suppress SiC MOSFET crosstalk?

Crosstalk（ crosstalk ）

chart 1

In the figure 1 In the half bridge circuit , The action tube is the lower tube S1, Applied to the upper tube S2 Is the off drive signal , Its body diode is in freewheeling state . When S1 When opening , Terminal voltage VDS1 falling , be S2 Start to withstand reverse voltage , The voltage at both ends VDS2 With dV/dt The speed of . that dV/dt Would pass S2 Of Crss Generate displacement current Irss=Crss*dV/dt,Irss Will flow S2 Drive circuit of , Yes CGS Charge , And in RG Pressure drop is generated on the . The end result is S2 The drive voltage of is pulled up , A positive spike appears , If exceeded S2 Of Vth, Will lead to misleading communication , Light will increase the loss , If the bridge arm is short circuited, it will explode . We call this process forward Crosstalk.

chart 2

S1 Still for the action tube , Only this time it is turned off . At this point, the whole process is the same as the forward Crosstalk The principle is the same , Only the voltage and current change in the opposite direction , Final S2 The drive voltage of is pulled down , There is a reverse spike . We all know SiC MOSFET The grid voltage withstand capability is very poor , Negative spikes can damage the grid , influence SiC MOSFET Or directly break down the grid . We call this process negative Crosstalk.

chart 3

chart 4

No matter Si MOSFET still IGBT All exist Crosstalk The problem of , Not at all SiC MOSFET Peculiar . however SiC MOSFET The switch is faster 、Vth Partial （ Generally in 2.5V-4.5V）、 Grid voltage withstand capability is weak , This makes Crosstalk Yes SiC MOSFET The consequences are even more serious 、 It is more difficult to deal with .

In order to suppress Crosstalk, The first thing to do is to measure accurately Crosstalk wave form . But for the following two reasons , Engineers often get wrong waveforms , It often leads to one pass operation as fierce as a tiger , The actual effect is 25 cents .

reason 1： Parasitic parameters affect

Just now I explained Crosstalk In principle , In order to express conciseness , chart 1 Sum graph 3 The circuit diagram given in is simplified . When considering many existing parasitic parameters , We get the graph 5 The equivalent circuit given in .SiC MOSFET The actual driving voltage on the chip is VGS, And what we get with the voltage probe is VGS-M. The difference between the two VGS-M Not just VGS, It also contains SiC MOSFET Chip gate resistance RG(int) Pressure drop across the VRG And parasitic inductance L Pressure drop across the VL. The reason for this is that the voltage probe cannot be directly connected to SiC MOSFET On the chip , It can only be connected to the pins of the device package , be RG(int) and L Between the measuring points .

chart 5

The simulation results show that , Measured by voltage probe Crosstalk The waveforms are better than what actually happens Crosstalk Low , That is to say , Due to the influence of parasitic parameters ,Crosstalk The severity of . This leads to the following two situations ： One is to judge by the measurement results Crosstalk To the extent acceptable , However, misleading communication has actually occurred ; Second, the engineers have made great efforts , It seems that Crosstalk Suppressed , In fact, it is still far away . because RG(int) and L Unavoidable , That is, the measurement error cannot be eliminated , Then the power engineer is using SiC MOSFET You need to be Crosstalk Leave enough margin .

meanwhile , The measurement results are consistent with the reality Crosstalk The difference between them will change with RG(int) and L To increase with , This enlightens us that we can choose RG(int) Of SiC MOSFET, At the same time, when measuring, try to connect the probe to the root of the device pin , This will minimize the error .

chart 6

chart 7

reason 2： Not using a suitable voltage probe

During power commissioning , The high-voltage differential probe is often used to measure the voltage signal , Its measuring range is wide 、 Differential input 、 The characteristic of high impedance is deeply loved by power engineering .

The picture shows Tektronix high pressure differential probe

But in measuring Crosstalk The waveform time difference probe is no longer applicable . First Crosstalk The range of amplitude is ±10V within , The attenuation multiple of high-pressure differential probe is large , This leads to large measurement error 、 It's noisy . secondly , The measuring line at the front end of the high-voltage differential probe is very long , Equivalent to an antenna , Will receive the SiC MOSFET Interference signal generated by rapidly changing current during switching , Thus affecting the measurement results . Last , The measuring line at the front end of the high-voltage differential probe can be regarded as inductance , It is easy to cause non-existent oscillation in the measurement results .

It can be seen from the measured results below , Measured with high-pressure differential probe Crosstalk The waveform is very thick , At the same time, its oscillation amplitude is very high , positive Crosstalk The peak has exceeded SiC MOSFET Of Vth（3.5V）, However, there is no misleading communication at this time , This shows that such test results are problematic , At the same time negative Crosstalk The peak has already surpassed SiC MOSFET Grid withstand voltage limit （-10V）. When the optical isolation probe is used Crosstalk The waveform is obviously different from that of the high-voltage differential probe , The wavy lines have become thinner , Both positive and negative Crosstalk The peaks are within the acceptable range . This is mainly due to the fact that the optical isolation probe can select a smaller attenuation multiple , At the same time, the connection between the front section of the probe and the device can realize the minimum loop connection .

chart 8

chart 9

The above is the measurement pipe Crosstalk Waveform of , So when we need to measure the upper tube Crosstalk When , What will happen . It can be seen from the measured waveform below , The waveform obtained by using the high-voltage differential probe is even more outrageous , The amplitude of its oscillation exceeds the positive 10V reverse 20V, The waveforms measured by the optical isolation probe are still within the acceptable range , This is mainly due to the excellent high frequency common mode rejection ratio of the optical isolation probe .

chart 10

chart 11

The picture shows Tektronix ISOVu A new generation of optical isolation probe

thus it can be seen , High pressure differential probes are not suitable for measurement Crosstalk, Getting the wrong waveform will mislead the circuit design , Waste engineers' time and energy . And selecting optical isolation probe can obtain accurate waveform , Whether measuring the lower pipe or the upper pipe , All have excellent performance .