The silicon carbide sand is crushed to the appropriate particle size, additives are added for batching, pressed and molded and dried at low temperature. After the moisture content reaches the required level, it is fired in a kiln. After the temperature is raised to 1450°C according to the corresponding temperature control requirements, it is properly insulated and then lowered to room temperature to obtain carbonization. Silicon products.
The production process of silicon carbide products: raw materials → ingredients → press molding (additives) → drying → firing → finished products. In the production process of domestic silicon carbide products, there are two production methods for the molding process of the products: manual shaping and press Made. The cost of manual production is low, but the product quality is unstable, the workload is large, and the operating environment is poor; the one-time investment of press production is large, but the product quality is stable, the labor demand is small, and the operating environment is good. Some large-scale domestic manufacturers have introduced presses to produce silicon carbide products in recent years.
In the production process of domestic silicon carbide products, there are two types of kilns commonly used in the firing process of the products: down-flame round kiln and down-flame square kiln.
The advantages of the round kiln are: the products are heated evenly when fired in the round kiln, the kiln body expands evenly after being heated, and the kiln body is durable, etc.; its disadvantage is that when firing rectangular silicon carbide trays, the circular inner cavity increases the volume of the kiln. The production capacity cannot be fully utilized, resulting in a reduction in production capacity. In the construction process of the circular kiln body, standard bricks cannot be used and specially processed wedge-shaped bricks must be used.
The advantages of the square kiln are: when firing rectangular silicon carbide trays, the internal volume of the kiln is fully utilized and the production capacity is large. Standard bricks are used in the kiln construction process, and the construction process is technically difficult. Its disadvantages are: the products are heated unevenly when fired in the square kiln, the kiln body expands unevenly after being heated, and the stress point is prone to deformation and damage to the kiln body.
Every engineer wants a perfect switch that switches instantaneously between on and off states with the lowest possible losses in both states. To achieve such switching characteristics, an infinite breakdown voltage is required, no current is allowed to flow when turned off, there is no need to maintain the above voltage difference when turned on, and the switching on and off occurs in an instant.
There is no such switch! The breakdown voltage of a real switch is limited, there is leakage current when it is turned off, and there is a voltage difference when it is turned on. Switching takes a certain amount of time and always consumes some energy.
Aren’t there now third-generation semiconductor devices such as silicon carbide (SiC)? However, although the characteristics of SiC devices are much better than those of silicon, there are still many new problems to be solved in applications.
Using SiC MOSFETs instead of silicon devices can increase the switching frequency by three to five times and achieve higher switching speeds by adjusting the driver stage to provide a higher gate-on voltage and handle the sometimes negative gate-off voltage. , while using smaller magnetics and heat sinks to save space.
However, as Zhao Zhengming, professor of the Department of Electrical Engineering and Applied Electronics Technology at Tsinghua University, doctoral supervisor, director of the Electronic Experiment Center of Tsinghua University, and IEEE Fellow, said: “The experience of applying SiC in the past 10 years is that doing it well is not as good as using it well. Switches If the speed is so fast, mutations will occur and cause many problems, so you will encounter great challenges when using SiC.” He believes that only by working together in all aspects of the supply chain can wide bandgap devices such as SiC be truly used.
Practical applications have shown that using SiC can significantly reduce turn-on losses, turn-off losses, and reverse recovery losses. The overall loss can be reduced by seven times, and the switching frequency can be greatly increased. In fact, original switches such as silicon-based MOSFETs
With IGBT, the switching frequency can also be increased, but the higher the frequency, the higher the loss, and the decrease in efficiency is of little significance.
He believes that the primary indicator of electric power conversion equipment is efficiency, and low efficiency has little value. Increasing the sampling frequency will make the waveform better. The operating frequency of SiC is almost 4 to 5 times higher than that of silicon-based devices, and the total loss is smaller, so efficiency can be improved. This is the main driver for the current use of SiC.
How much benefit can the application actually have?
High voltage is one of the important characteristics of SiC. Originally, in order to achieve high-voltage applications, single silicon devices could only be connected in series or pieced together in cascade to withstand high voltages. Since SiC has a high withstand voltage, it can be used directly, bringing benefits to various applications. Comes great convenience.
From the development of high-voltage SiC chip technology, there are two combinations of 3.3kV high-voltage SiC devices currently available on the market:
SiC MOSFET is used, and the feedback diode uses the SiC MOSFET body diode.
Use a single MOSFET plus an anti-parallel diode.
Both options have their own benefits. Over time, the forward voltage drop of the SiC MOSFET body diode will change. The 3.3kV high-voltage SiC device sold by Mitsubishi Electric has a separate parallel feedback diode on the outside, which can avoid the forward degradation of the diode; the next-generation 3.3kV full SiC module integrates this diode into the chip and makes it into a chip, which can avoid After long-term operation, the forward voltage drop of the MOSFET body diode changes, which can also increase the power density of the module.
The currently mass-produced 3.3kV 750A SiC MOSFET has a built-in diode chip (SBD). The 3.3kV product has two insulation levels: 6kV (standard LV100 package) and 10kV (standard HV100 package). The 6.5kV SiC MOSFET under development also has built-in SBD, which greatly reduces the chip area and the power density reaches 9.3kVA/cm3. It uses high thermal conductivity and high heat resistance insulating substrate and high reliability chip welding technology. The chip has Good heat dissipation and heat resistance. By increasing the switching frequency of the device, it helps to achieve miniaturization and energy saving of high-voltage converters;
According to reports, high-voltage SiC modules have become the mainstream of applications. One of the main applications of 3.3kV is rectifiers and inverters in rail traction.
Application on trains; In 2019, the 3.3kV 750A full SiC MOSFET power module in LV100 package was applied on Odakyu 5000 trains. Today, rail traction at home and abroad no longer uses the previous diode rectifier, but uses SiC MOSFET and IGBT devices.
The second application is medium and low voltage power transmission and distribution. In order to improve efficiency, SiC devices need to be used to achieve medium and low voltage power transmission, generally using MMC structures. Among them, power electronics
The transformer adopts full-bridge and half-bridge topology, including rectifier part, isolation part and inverter part. Generally, SiC devices are mostly used in the isolation part, which can effectively reduce the size of the transformer.
In terms of power density, it is developing from the past IGBT technology through packaging optimization to modules using SiC MOSFET. The earliest H series power density was only 4.51 cm2, and now it is 10.71A/cm2, which has increased by more than twice, making the entire converter The volume and efficiency of the device have been qualitatively optimized.
Among the mass-produced products, there are 3.3kV hybrid and all-carbon devices. The hybrid includes the IGBT part and the diode is all SiC. The advantage is that most customers can accept this cost and the loss can be reduced a lot. Because the diode is made of SiC material, the reverse recovery loss
Very low inhibition. In some areas, such as applications where diodes dominate, this module is ideally suited. Current products are available in two grades: 1700V 1200A and 3.3kV 600A.
Includes three models: 750A, 375A, 185A. 185A is being tested. Using this device, the volume and efficiency of the entire converter will be greatly optimized. Its packaging is the same as that of silicon devices, but the structural design and layout can be simplified a lot.
Comparing the turn-on waveforms of 3.3kV 600A silicon devices, hybrid SiC devices and full SiC devices, it can be seen that the turn-on loss of silicon devices is 1.02, hybrid SiC devices are 0.62, and SiC devices are 0.42. Full SiC is about 60% lower than silicon, and the effect is very obvious.
In terms of turn-off loss, we mainly compared silicon and hybrid SiC. Because their front ends are both IGBTs, the losses are relatively large. After switching to SiC devices, the reduction dropped from 1.05 to 0.14, which is a very large reduction.
Power consumption simulation is very illustrative of the problem. The simulation conditions are: bus voltage 1800V, current level 450A, frequency 5kHz. This is the optimal frequency for 3.3kV full SiC devices, which is about 10 times higher than the previous silicon devices. In the past, 3.3kV was used for traction. Silicon devices are below 1kHz. In terms of overall loss, all-SiC devices are 76% lower than silicon devices. As conditions change, this loss will fluctuate, but the reduction is still very large.
How to resolve application pain points?
The driver of high-voltage SiC modules is a difficult problem in application. Perhaps it is because Mitsubishi Electric is a major power semiconductor manufacturer and does not develop its own drivers. Instead, many manufacturers develop drivers specifically for it. Several of its recommended drives have been approved after testing, including Japanese brands IDC and Tamura, American PI (Power Integrations) and domestic brand Bronze Sword.
The domestic application of 3.3kV high-voltage SiC modules has just begun. In order to deal with some problems encountered in product development practice or production, Mitsubishi Electric has developed a process component
design. The module used in its topology is a 3.3kV 750A SiC module. Because the current level for traction applications is slightly smaller, parallel connection is very necessary. This component is connected in parallel with two 3.3kV 750A SiC modules.
This platform can be used to conduct specific research on the driving characteristics of SiC and provide it to customers for testing, to reproduce laboratory test results and solve some practical problems in applications; it can also evaluate some driver boards with the same package. Its development goals are rated power 1.2MW, output voltage 1140Vrms, output current 600Arms, bus voltage 1800V to 2200V, and frequency setting 5kHz. Because it is a parallel application, the key point is that the current imbalance rate is less than 10%. This component will be used in rail traction, DC power transmission and distribution, high-voltage frequency conversion and other fields in the future.
It includes two SiC power modules and customized busbars. The driver solution uses Japanese products, including the main board and slave board; there are also customized DC side capacitors and radiators.
Many problems were encountered in the design of this component, especially the drive design and transient commutation circuit design, as well as parallel connection issues. Since the driver design uses mature drivers, there are no problems. The main concern is the transient commutation loop
design, which is different from the previous ordinary circuit design. If the frequency of ordinary devices is relatively low, the parasitic parameters of the devices in the circuit are rarely considered. However, the transient commutation circuit must consider the transient parasitic parameters, such as stray inductance, stray capacitance, etc., because these will affect the turn-on of devices connected in parallel with each other. and turn-off transient behavior, thus affecting the performance of IGBT or MOSFET. Research has found that parasitic inductance is the most important in high-power design, so parasitic inductance should be regarded as a key target.
The design of the transient commutation loop is to consider the influence of stray parameters on the commutation process, so the stray parameters are taken as the design target of the transient commutation loop. The transient commutation loop includes two parts: the system commutation loop and the unit commutation loop.
The goal of the system commutation loop is to control the stray inductance in the loop, because stray inductance can cause very high voltage spikes during the commutation process. If it is too high, it will breakdown the IGBT module; the unit commutation loop is to solve the problem of module turn-on And the problem of inconsistent turn-off time, because the electrical connections of the modules are the same, but the structures are different. The busbars and wire lengths of module 1 (red) and module 2 (green) are different, and many of the parasitic parameters will have some differences (yellow in the picture) circle), it will cause the module turn-on and turn-off times to be inconsistent, causing current sharing problems. Therefore, the design goal of the unit commutation circuit is to ensure that the stray parameters of the parallel power modules are as close as possible, so that the current change rate of the control module is also as close as possible to achieve a good current sharing effect.
The design of the transient commutation loop consists of three steps:
first step
Transient commutation loop system analysis, including extraction of transient commutation loop; judging stray parameters, including capacitance, inductance, etc., to see which parameter is the most important to the system, taking the most important one as the design goal, and then simplifying the circuit, in the plan Include only the most important design goals; determine the range within which spurious parameters should be controlled based on system requirements.
Step 2
Carry out theoretical design, including theoretical calculation and system modeling and simulation, to optimize the busbar and structural layout, and finally attribute the stray parameters to the structural design.
third step
Experimental verification of the transient commutation loop, optimizing the busbar, and conducting experiments after adjusting the layout. Mainly test stray parameters and compare them with the previous goals to see if the requirements are met; verify the safe working area and see if the modules are turned on and off. Check whether the fault is within the safe working area.
According to the design plan, the system commutation circuit design actually takes the busbar as the design target. In this plan, the busbar inductance target is set to less than 20nH, so that the shutdown voltage can be controlled within the safe working area of about 2200V. to meet system requirements. The unit commutation circuit is to make some parasitic parameters consistent after module 1 and module 2 are extracted, thereby making the opening and closing consistent.
After the previous designs are completed, verify the entire system and select a driver for verification.
Mainly for parallel test verification, select three test conditions: turn on 750A, turn off 750A (both IGBT ratings), and turn off 1500A (twice the rated current). It can be seen that the current sharing efficiency
Dividing the difference between the two currents by an average gives 1.6%, 0.9% and 1.1%. The peak voltage is controlled at 2200V, and the double rated current shutdown can also be controlled at 2300V. The effect is very good, and the current sharing and voltage are well controlled.
As Professor Zhao Zhengming said, fast short-circuit protection is the biggest reliability issue in the use of SiC modules. It is affected by many factors, including the body diode and reverse bias diode of the SiC chip itself, which are prone to problems, causing gate interference or parameters. Mismatch. In addition, due to the high switching speed, high electrical stress is easily generated between the DS and GS pins, making the pulse very high; commercial SiC modules are not very mature in terms of packaging technology, including welding. In addition, because the device is too fast, the fault protection speed is not enough and there is no time to respond.
Therefore, the short-circuit characteristics of SiC modules are relatively strict, and each manufacturer has proposed various protection measures. Mitsubishi Electric’s modules perform two parallel short-circuit tests.
The silicon carbide sand is crushed to the appropriate particle size, additives are added for batching, pressed and molded and dried at low temperature. After the moisture content reaches the required level, it is fired in a kiln. After the temperature is raised to 1450°C according to the corresponding temperature control requirements, it is properly insulated and then lowered to room temperature to obtain carbonization. Silicon products.
The production process of silicon carbide products: raw materials → ingredients → press molding (additives) → drying → firing → finished products. In the production process of domestic silicon carbide products, there are two production methods for the molding process of the products: manual shaping and press Made. The cost of manual production is low, but the product quality is unstable, the workload is large, and the operating environment is poor; the one-time investment of press production is large, but the product quality is stable, the labor demand is small, and the operating environment is good. Some large-scale domestic manufacturers have introduced presses to produce silicon carbide products in recent years.
In the production process of domestic silicon carbide products, there are two types of kilns commonly used in the firing process of the products: down-flame round kiln and down-flame square kiln.
The advantages of the round kiln are: the products are heated evenly when fired in the round kiln, the kiln body expands evenly after being heated, and the kiln body is durable, etc.; its disadvantage is that when firing rectangular silicon carbide trays, the circular inner cavity increases the volume of the kiln. The production capacity cannot be fully utilized, resulting in a reduction in production capacity. In the construction process of the circular kiln body, standard bricks cannot be used and specially processed wedge-shaped bricks must be used.
The advantages of the square kiln are: when firing rectangular silicon carbide trays, the internal volume of the kiln is fully utilized and the production capacity is large. Standard bricks are used in the kiln construction process, and the construction process is technically difficult. Its disadvantages are: the products are heated unevenly when fired in the square kiln, the kiln body expands unevenly after being heated, and the stress point is prone to deformation and damage to the kiln body.
Every engineer wants a perfect switch that switches instantaneously between on and off states with the lowest possible losses in both states. To achieve such switching characteristics, an infinite breakdown voltage is required, no current is allowed to flow when turned off, there is no need to maintain the above voltage difference when turned on, and the switching on and off occurs in an instant.
There is no such switch! The breakdown voltage of a real switch is limited, there is leakage current when it is turned off, and there is a voltage difference when it is turned on. Switching takes a certain amount of time and always consumes some energy.
Aren’t there now third-generation semiconductor devices such as silicon carbide (SiC)? However, although the characteristics of SiC devices are much better than those of silicon, there are still many new problems to be solved in applications.
Using SiC MOSFETs instead of silicon devices can increase the switching frequency by three to five times and achieve higher switching speeds by adjusting the driver stage to provide a higher gate-on voltage and handle the sometimes negative gate-off voltage. , while using smaller magnetics and heat sinks to save space.
However, as Zhao Zhengming, professor of the Department of Electrical Engineering and Applied Electronics Technology at Tsinghua University, doctoral supervisor, director of the Electronic Experiment Center of Tsinghua University, and IEEE Fellow, said: “The experience of applying SiC in the past 10 years is that doing it well is not as good as using it well. Switches If the speed is so fast, mutations will occur and cause many problems, so you will encounter great challenges when using SiC.” He believes that only by working together in all aspects of the supply chain can wide bandgap devices such as SiC be truly used.
Practical applications have shown that using SiC can significantly reduce turn-on losses, turn-off losses, and reverse recovery losses. The overall loss can be reduced by seven times, and the switching frequency can be greatly increased. In fact, original switches such as silicon-based MOSFETs
With IGBT, the switching frequency can also be increased, but the higher the frequency, the higher the loss, and the decrease in efficiency is of little significance.
He believes that the primary indicator of electric power conversion equipment is efficiency, and low efficiency has little value. Increasing the sampling frequency will make the waveform better. The operating frequency of SiC is almost 4 to 5 times higher than that of silicon-based devices, and the total loss is smaller, so efficiency can be improved. This is the main driver for the current use of SiC.
How much benefit can the application actually have?
High voltage is one of the important characteristics of SiC. Originally, in order to achieve high-voltage applications, single silicon devices could only be connected in series or pieced together in cascade to withstand high voltages. Since SiC has a high withstand voltage, it can be used directly, bringing benefits to various applications. Comes great convenience.
From the development of high-voltage SiC chip technology, there are two combinations of 3.3kV high-voltage SiC devices currently available on the market:
SiC MOSFET is used, and the feedback diode uses the SiC MOSFET body diode.
Use a single MOSFET plus an anti-parallel diode.
Both options have their own benefits. Over time, the forward voltage drop of the SiC MOSFET body diode will change. The 3.3kV high-voltage SiC device sold by Mitsubishi Electric has a separate parallel feedback diode on the outside, which can avoid the forward degradation of the diode; the next-generation 3.3kV full SiC module integrates this diode into the chip and makes it into a chip, which can avoid After long-term operation, the forward voltage drop of the MOSFET body diode changes, which can also increase the power density of the module.
The currently mass-produced 3.3kV 750A SiC MOSFET has a built-in diode chip (SBD). The 3.3kV product has two insulation levels: 6kV (standard LV100 package) and 10kV (standard HV100 package). The 6.5kV SiC MOSFET under development also has built-in SBD, which greatly reduces the chip area and the power density reaches 9.3kVA/cm3. It uses high thermal conductivity and high heat resistance insulating substrate and high reliability chip welding technology. The chip has Good heat dissipation and heat resistance. By increasing the switching frequency of the device, it helps to achieve miniaturization and energy saving of high-voltage converters;
According to reports, high-voltage SiC modules have become the mainstream of applications. One of the main applications of 3.3kV is rectifiers and inverters in rail traction.
Application on trains; In 2019, the 3.3kV 750A full SiC MOSFET power module in LV100 package was applied on Odakyu 5000 trains. Today, rail traction at home and abroad no longer uses the previous diode rectifier, but uses SiC MOSFET and IGBT devices.
The second application is medium and low voltage power transmission and distribution. In order to improve efficiency, SiC devices need to be used to achieve medium and low voltage power transmission, generally using MMC structures. Among them, power electronics
The transformer adopts full-bridge and half-bridge topology, including rectifier part, isolation part and inverter part. Generally, SiC devices are mostly used in the isolation part, which can effectively reduce the size of the transformer.
In terms of power density, it is developing from the past IGBT technology through packaging optimization to modules using SiC MOSFET. The earliest H series power density was only 4.51 cm2, and now it is 10.71A/cm2, which has increased by more than twice, making the entire converter The volume and efficiency of the device have been qualitatively optimized.
Among the mass-produced products, there are 3.3kV hybrid and all-carbon devices. The hybrid includes the IGBT part and the diode is all SiC. The advantage is that most customers can accept this cost and the loss can be reduced a lot. Because the diode is made of SiC material, the reverse recovery loss
Very low inhibition. In some areas, such as applications where diodes dominate, this module is ideally suited. Current products are available in two grades: 1700V 1200A and 3.3kV 600A.
Includes three models: 750A, 375A, 185A. 185A is being tested. Using this device, the volume and efficiency of the entire converter will be greatly optimized. Its packaging is the same as that of silicon devices, but the structural design and layout can be simplified a lot.
Comparing the turn-on waveforms of 3.3kV 600A silicon devices, hybrid SiC devices and full SiC devices, it can be seen that the turn-on loss of silicon devices is 1.02, hybrid SiC devices are 0.62, and SiC devices are 0.42. Full SiC is about 60% lower than silicon, and the effect is very obvious.
In terms of turn-off loss, we mainly compared silicon and hybrid SiC. Because their front ends are both IGBTs, the losses are relatively large. After switching to SiC devices, the reduction dropped from 1.05 to 0.14, which is a very large reduction.
Power consumption simulation is very illustrative of the problem. The simulation conditions are: bus voltage 1800V, current level 450A, frequency 5kHz. This is the optimal frequency for 3.3kV full SiC devices, which is about 10 times higher than the previous silicon devices. In the past, 3.3kV was used for traction. Silicon devices are below 1kHz. In terms of overall loss, all-SiC devices are 76% lower than silicon devices. As conditions change, this loss will fluctuate, but the reduction is still very large.
How to resolve application pain points?
The driver of high-voltage SiC modules is a difficult problem in application. Perhaps it is because Mitsubishi Electric is a major power semiconductor manufacturer and does not develop its own drivers. Instead, many manufacturers develop drivers specifically for it. Several of its recommended drives have been approved after testing, including Japanese brands IDC and Tamura, American PI (Power Integrations) and domestic brand Bronze Sword.
The domestic application of 3.3kV high-voltage SiC modules has just begun. In order to deal with some problems encountered in product development practice or production, Mitsubishi Electric has developed a process component
design. The module used in its topology is a 3.3kV 750A SiC module. Because the current level for traction applications is slightly smaller, parallel connection is very necessary. This component is connected in parallel with two 3.3kV 750A SiC modules.
This platform can be used to conduct specific research on the driving characteristics of SiC and provide it to customers for testing, to reproduce laboratory test results and solve some practical problems in applications; it can also evaluate some driver boards with the same package. Its development goals are rated power 1.2MW, output voltage 1140Vrms, output current 600Arms, bus voltage 1800V to 2200V, and frequency setting 5kHz. Because it is a parallel application, the key point is that the current imbalance rate is less than 10%. This component will be used in rail traction, DC power transmission and distribution, high-voltage frequency conversion and other fields in the future.
It includes two SiC power modules and customized busbars. The driver solution uses Japanese products, including the main board and slave board; there are also customized DC side capacitors and radiators.
Many problems were encountered in the design of this component, especially the drive design and transient commutation circuit design, as well as parallel connection issues. Since the driver design uses mature drivers, there are no problems. The main concern is the transient commutation loop
design, which is different from the previous ordinary circuit design. If the frequency of ordinary devices is relatively low, the parasitic parameters of the devices in the circuit are rarely considered. However, the transient commutation circuit must consider the transient parasitic parameters, such as stray inductance, stray capacitance, etc., because these will affect the turn-on of devices connected in parallel with each other. and turn-off transient behavior, thus affecting the performance of IGBT or MOSFET. Research has found that parasitic inductance is the most important in high-power design, so parasitic inductance should be regarded as a key target.
The design of the transient commutation loop is to consider the influence of stray parameters on the commutation process, so the stray parameters are taken as the design target of the transient commutation loop. The transient commutation loop includes two parts: the system commutation loop and the unit commutation loop.
The goal of the system commutation loop is to control the stray inductance in the loop, because stray inductance can cause very high voltage spikes during the commutation process. If it is too high, it will breakdown the IGBT module; the unit commutation loop is to solve the problem of module turn-on And the problem of inconsistent turn-off time, because the electrical connections of the modules are the same, but the structures are different. The busbars and wire lengths of module 1 (red) and module 2 (green) are different, and many of the parasitic parameters will have some differences (yellow in the picture) circle), it will cause the module turn-on and turn-off times to be inconsistent, causing current sharing problems. Therefore, the design goal of the unit commutation circuit is to ensure that the stray parameters of the parallel power modules are as close as possible, so that the current change rate of the control module is also as close as possible to achieve a good current sharing effect.
The design of the transient commutation loop consists of three steps:
first step
Transient commutation loop system analysis, including extraction of transient commutation loop; judging stray parameters, including capacitance, inductance, etc., to see which parameter is the most important to the system, taking the most important one as the design goal, and then simplifying the circuit, in the plan Include only the most important design goals; determine the range within which spurious parameters should be controlled based on system requirements.
Step 2
Carry out theoretical design, including theoretical calculation and system modeling and simulation, to optimize the busbar and structural layout, and finally attribute the stray parameters to the structural design.
third step
Experimental verification of the transient commutation loop, optimizing the busbar, and conducting experiments after adjusting the layout. Mainly test stray parameters and compare them with the previous goals to see if the requirements are met; verify the safe working area and see if the modules are turned on and off. Check whether the fault is within the safe working area.
According to the design plan, the system commutation circuit design actually takes the busbar as the design target. In this plan, the busbar inductance target is set to less than 20nH, so that the shutdown voltage can be controlled within the safe working area of about 2200V. to meet system requirements. The unit commutation circuit is to make some parasitic parameters consistent after module 1 and module 2 are extracted, thereby making the opening and closing consistent.
After the previous designs are completed, verify the entire system and select a driver for verification.
Mainly for parallel test verification, select three test conditions: turn on 750A, turn off 750A (both IGBT ratings), and turn off 1500A (twice the rated current). It can be seen that the current sharing efficiency
Dividing the difference between the two currents by an average gives 1.6%, 0.9% and 1.1%. The peak voltage is controlled at 2200V, and the double rated current shutdown can also be controlled at 2300V. The effect is very good, and the current sharing and voltage are well controlled.Silicon carbid.
As Professor Zhao Zhengming said, fast short-circuit protection is the biggest reliability issue in the use of SiC modules. It is affected by many factors, including the body diode and reverse bias diode of the SiC chip itself, which are prone to problems, causing gate interference or parameters. Mismatch. In addition, due to the high switching speed, high electrical stress is easily generated between the DS and GS pins, making the pulse very high; commercial SiC modules are not very mature in terms of packaging technology, including welding. In addition, because the device is too fast, the fault protection speed is not enough and there is no time to respond.
Therefore, the short-circuit characteristics of SiC modules are relatively strict, and each manufacturer has proposed various protection measures. Mitsubishi Electric’s modules perform two parallel short-circuit tests.