Advertisement



Cybertruck Power Electronics

TruckElectric

Well-known member
First Name
Bryan
Joined
Jun 16, 2020
Messages
445
Reaction score
545
Location
Texas
Vehicles
Dodge Ram diesel
Occupation
Retired
Country flag
If I'm not mistaken the Cybertruck is going to use the same motors as the Model 3. The Power Electronics/Power Module will most definitely be upgraded for the Cybertruck.

It will interesting to get a glimpse of the PE once it becomes available.

Here is an article on the Model 3 Power Module:


Here’s why Tesla transitioned to a semi-custom power module design in Model 3 inverter

A closer look at semiconductor packaging considerations in EVs

One of the most critical decisions to be made at the earliest stage of designing a new power converter concerns the packages used for the semiconductors, as pretty much every other aspect of the design hinges on their physical form. This is especially true for the main power converters used in EVs—on-board chargers, DC/DC converters and inverters—as there are tight constraints on the size (and cost, of course) allotted to each. Furthermore, any device that has direct or incidental contact with the AC mains will also need to meet some rather onerous electrical safety requirements which—as a case study below will show—can critically depend on the package used for the semiconductor switches.


The power semiconductor components most likely to be used in EVs come in two different form factors: (1) plastic types such as the TO-220 and TO-247 packages, which feature wire leads and a (usually non-isolated) heatsink tab, and which typically contain a single diode or switch (with or without anti-parallel diode); (2) modules, which typically contain several components pre-wired in commonly used configurations (e.g. a half-bridge plus a temperature sensor), all mounted on an electrically-isolated heat spreader. Modules also tend to have screw terminals for the high-power connections and pin or spring terminals for the low-power connections, making integration into a bused structure (and replacement of a damaged module) much easier. Despite the radical differences in their physical (and, often, electrical) aspects, there’s no clear distinction for when to choose a plastic package component or a module; using a rather broad brush to delineate between the two, modules are preferred if more than 50-100 A RMS must be handled, whereas plastic packages are preferred if switching frequency must be considerably above the ultrasonic range (e.g. >40 kHz). These are obviously very different criteria, nor are they mutually exclusive, but suffice it to say that if you need to switch >100 A RMS at >100 kHz, then you’re looking at a design challenge worthy of a PhD dissertation.

4-Semiconductor-packaging-considerations.png



There are numerous other criteria as well as exceptions to the above rules of thumb—for a notable example, Tesla was quite fond of using dozens (84!) of TO-247 switches in its earlier inverters—but it is telling that the Model 3 inverter uses what might be called a quasi-modular approach, with far fewer devices (24) of much higher individual power rating, but still in a plastic package-like form. In fact, why Tesla might have chosen a TO-247 package device at first, only to transition to a semi-custom module later on, is precisely the subject of this article.

at-semiconductor-packaging-considerations-in-EVs-1.png


Prior to the emergence of OEM EVs, power semiconductor modules were designed specifically for industrial applications, with the vast majority being used in 3-phase motor drives supplied by the AC mains. Consequently, the available voltage ratings were in rather coarse steps of 600 V for 208-240 VAC mains applications, 1,200 V for 440-480 VAC, 1,700 V for 575-600 VAC, and so on. Furthermore, 3-phase motors also come in rather coarse power rating steps, so the current ratings for modules were equally coarse as well. Also, industrial applications tend to be more concerned with reliability and efficiency than with minimizing size (and the noise from “singing” motors and transformers), so the diodes and switches inside the modules weren’t particularly fast (i.e. PWM frequency rarely exceeded 10 kHz, and was usually closer to 1 kHz, especially at 1,200 V and above). Finally, while the market for all industrial motor drives is quite large, the market for any one particular voltage/current combination is relatively small, and some combinations of voltage/current just don’t make sense industrially. For example, it is possible to get 1,700 V modules rated for 3,500 A or higher, but for 600 V modules the highest current rating commonly available is 600 A. This is because no (sane) industrial customer is going to try running a >200 hp motor from 240 V mains!

k-at-semiconductor-packaging-considerations-in-EVs.png


2-Semiconductor-packaging-considerations.png


Conversely, there is a veritable smorgasbord of devices and ratings in the plastic TO-247 (and smaller TO-220) packages, such that any practical combination of voltage, current and switching frequency can be had with judicious circuit design and layout (and a PCB capable of handling the current). More specifically, almost any current rating can be obtained by paralleling as many TO-247 devices as necessary…at the rate of about 20 A to 50 A per device, depending on the device technology, total losses, and how heat from said losses is removed from the junction (but note that it takes increasingly heroic measures to keep the junction temperature of a TO-247 device below 100-125° C once dissipation exceeds 50 W). For example, SiC MOSFETs have extremely low switching and conduction losses, and can tolerate operation at much higher temperatures than conventional Si MOSFETs or IGBTs, so the limiting factor on how much current can be crammed through one in a TO-247 package might very well be the ampacity of the bond-wires and/or leads. In contrast, a TO-247 IGBT with a fairly constant voltage drop of 2.2 V and comparatively high switching losses might struggle to handle 25 A, even with liquid cooling. Another factor that greatly affects the ampacity per device is that the heatsink tab on the conventional TO-247 and TO-220 packages is directly connected to the collector or drain, for IGBTs and MOSFETs, respectively, so some form of insulator will be needed between the tab and the heatsink. Unfortunately, most materials which are good electrical insulators are also good thermal insulators, such that even extremely thin sheets of mica, silicone rubber or Kapton (aka polyimide) will add around 1° C/W of thermal resistance to a TO-247 package (and up to 3° C/W for the smaller TO-220). This resistance adds to that of the junction to case and the heatsink to ambient pathways, hence the practical upper limit of 50 W dissipation per TO-247.

3-Semiconductor-packaging-considerations.png



There are a couple of exceptions to the “good electrical insulator = poor thermal conductor” rule: aluminum oxide and nitride. The former has a bulk thermal conductivity of 30 W/m-K, while the latter clocks in at 285 W/m-K [see sidebar below: Thermal conductivity vs resistance]. Both compare rather favorably to the thermal conductivity of mica at 0.3 W/m-K (or 100x to almost 1,000x worse), but aluminum nitride is an even better conductor of heat than pure aluminum (235 W/m-K), though still not as good as pure copper (400 W/m-K). Both aluminum oxide and nitride are ceramic-like materials that are hard and brittle, and also like ceramics, they are refractory (i.e. they have a very high melting point), so insulators made from them have to be relatively thick (1 mm seems to be a practical limit) compared to mica (~0.1 to 0.3 mm) or silicone rubber (<0.5 mm). Even so, aluminum oxide and nitride insulators are quite fragile. For example, a product I have helped to redesign utilizes SiC MOSFETs in a TO-247 package with 1 mm-thick aluminum nitride insulators between them and the extruded aluminum heatsink. During UL “open/short” testing (in which the UL inspector randomly opens or shorts various components, looking for potential safety issues), one of the switches exploded and shattered the aluminum nitride insulator. This allowed excessive fault current into earth ground, which is a definite fail of the test (it is perfectly acceptable for your product to quit working during this particular test, it just can’t catch fire or create a shock hazard). Changing the fuse to a faster-acting type (read: more expensive, and more prone to “nuisance trip”) sufficiently limited fault energy to less than what is needed to rupture a TO-247 package, but this is not the sort of thing you want to deal with at the proverbial eleventh hour.

k-at-semiconductor-packaging-considerations-in-EVs.png


4-A-closer-look-at-semiconductor-packaging-considerations-in-EVs.png


This discussion circuitously segues back to a considerable advantage of modules: the heatsink “tab” is already electrically insulated from the semiconductor dice, and the dice themselves are typically encapsulated in a special silicone gel, which both improves heat removal from the bondwires and does a decent job of containing shrapnel and metal vapor should things go pear-shaped. More specifically, the usual construction of a module is a sandwich consisting of dice soldered to intermediate heat spreaders (usually of copper) to increase the area available for transferring heat (and provide a common electrical connection between dice), followed by an aluminum oxide or nitride sheet which provides electrical isolation and, finally, a single heat spreader which also serves as the mounting baseplate. Basically, the intermediate heat spreaders lower the total thermal resistance from junction to heatsink compared to a solution using multiple TO-247 components, while the silicone gel and aluminum nitride insulators provide considerable voltage withstand rating. In fact, most (if not all) modules for industrial applications have been “recognized” by the major safety agencies (UL, TUV, Intertek, etc) for a given voltage withstand (or “hi-pot”) rating, which makes passing their tests a lot easier (by taking less time and costing less money).

k-at-semiconductor-packaging-considerations-in-EVs.png


As is usually the case, there are advantages and disadvantages to both types of packaging technology for semiconductors—there’s no such thing as a “one size fits all solution”—and so, unsurprisingly, none of the existing offerings are ideally suited to EV (or hybrid) applications. This is where the advanced packaging solutions conjured up for the Model 3 by Tesla and STMicroelectronics come in: a new approach to module design that combines the low-cost and reduced stray inductance of a plastic package with the electrical isolation, improved thermal performance and greater current rating per device of a classic industrial module, all with a form factor tailored for EV applications, rather than the “one size sort of fits most” of yesterday’s technology.

Sidebar: Thermal conductivity vs resistance

These two specs get tossed around a lot in power electronics, and while thermal resistance is easy to grasp—it is the thermal analog to electrical resistance, with heat in watts as current (okay, that is admittedly confusing) and temperature rise as voltage—thermal conductivity seems to be a bit trickier, mainly because of the confusing units it is given in of W/m-K (or in Imperial units of…on second thought, let’s not even go there). The key difference is that thermal conductivity describes a property of a material in general, while thermal resistance describes a specific use of that material. For example, a 200 mm2 * 1 mm thick insulator pad made of alumina (i.e. aluminum oxide, which has a thermal conductivity of 30 W/m-K) will have a thermal resistance of 0.167° C/W. If the thickness is doubled to 2 mm, then the thermal resistance will also double, while if the area is doubled then the thermal resistance will be halved. In all cases the thermal conductivity is the same, however. Similarly, if the same 200 mm2 * 1 mm insulator is made out of aluminum nitride, then its thermal resistance will plummet to less than 0.018° C/W, or 9.5x less, which makes sense given that aluminum nitride has a thermal conductivity 9.5x better than aluminum oxide.


Source: CHARGED ELECTRIC VEHICLES MAGAZINE
 
Last edited:

Geo

Well-known member
First Name
George
Joined
Mar 29, 2020
Messages
174
Reaction score
180
Location
Melbourne, Australia
Vehicles
Subaru STI, KTM450EXC
Country flag
In my best DiCaprio impersonation, of his impersonation of Frank Abignale : " I concur "

AjDelange, perhaps you have something more constructive to add ;)
 

ajdelange

Well-known member
First Name
A. J.
Joined
Dec 8, 2019
Messages
1,567
Reaction score
1,426
Location
Virginia/Quebec
Vehicles
Tesla X LR+, Lexus SUV, Toyota SR5, Toyota Landcruiser
Occupation
EE (Retired)
Country flag
My most recent experience in "automotive electronics" was received 60 years ago when I had a couple of summer jobs designing rectifiers some of which were for the Saginaw Steering Gear Division of General Motors who used them to chrome plate bumpers. Things have come along a bit since then. For example, our "switches" were SCRs which can be gated on but not off.
 
OP
TruckElectric

TruckElectric

Well-known member
First Name
Bryan
Joined
Jun 16, 2020
Messages
445
Reaction score
545
Location
Texas
Vehicles
Dodge Ram diesel
Occupation
Retired
Country flag
  • Thread starter
  • Thread Starter
  • #7
Silicon Carbide
breakthroughs to
accelerate electric
vehicle innovation

Anup Bhalla explores how wide band-gap SiC can help spur EV developments

Consumer demand for electric vehicles (EVs) with a range comparable to internal combustion engines is currently outrunning the technology itself and, crucially, at a more accessible price point. More efficient drivetrains using semiconductor technologies such as Silicon Carbide (SiC) are enabling engineers to achieve the high voltage and power demands in a cost-effective way.
Electric cars are becoming more mainstream, with prices coming down and range going up. Sales of electric cars topped 2.1 million globally in 2019. According to the International Energy Agency’s report ‘Global EV Outlook
page1image46684160

2020’, there were over 7.2 million electric passenger cars on the roads in 2019. However, continued growth depends on a number of factors. The reduced purchase subsidies in key markets has contributed to a significant drop in sales. There is also the COVID-19 pandemic to consider, which has had a major impact on car production globally this year. However, charging infrastructure improvements and consumer expectations of further technology improvements to lower vehicle prices and importantly, to give better range, remain key challenges.
"
More efficient drivetrains using semiconductor technologies such as Silicon Carbide (SiC) are enabling engineers to achieve the high voltage and power demands in a cost-effective way
Battery and motor manufacturers are reaching physical limits for performance using known technologies. However, in the drivetrain, where battery energy is converted into three-phase AC power for the motors, there is an obvious upgrade path away from traditional designs. This is to use wide band-gap semiconductors such as SiC.

Some EV applications have already started using SiC technology. These have largely been for low power
applications, such as battery chargers, auxiliary DC-DC converters and solid-state circuit breakers. However, drivetrain power designers have been reluctant to use this technology, preferring to wait until it could achieve an acceptably low ON-resistance, better robustness and easier application. Now, a breakthrough in performance is addressing all of those concerns—the latest generation SiC- FET or ‘stacked cascode’ from UnitedSiC.

What is a stacked cascode?

A stacked cascode is a device with two transistors stacked on top of one another: a high-voltage SiC JFET is connected in series with an optimised low voltage Si-MOSFET (see Figure 1). When the gate is high the MOSFET shorts the JFET gate-source, turning it ON. When the gate is low, the MOSFET drain voltage rises but only to the point that the JFET pinches OFF (ie the channel closes), which is around 10V. The result is a normally-OFF device with an easy gate drive. Moreover, it has all the benefits of a SiC device with low ON-resistance, high voltage and high-temperature operation, and an integral body diode effect with excellent reverse recovery characteristics.
page4image36433280

Figure 1: Stacked cascode construction – the MOSFET die is physically stacked on top of the JFET source pad

The cascode idea has been around for some time now but JFET versions are now achieving ON-resistances at high voltage ratings, making them close to the ‘ideal’ switch. Putting some figures to it, Table 1 is a selection of SiC-FETs from UnitedSiC showing RDS(ON) figures as low as 8.6 milliohms for a 1200V device and 6.7 Milliohms for a 650V device, both at 25 degrees Celsius. All devices are in the TO-247 package format, some with 4-lead Kelvin connections for optimum gate drive.

Table-1.jpg

Table 1: Latest generation UnitedSiC SiC-FET performance

Low drain-source on resistance (RDS(ON)), output capacitance (COSS) and switching energy (EON and EOFF) values reduce conduction losses to a minimum. In
page4image36433488

addition, switches with inductive loads, such as in motor drives, have to ‘commutate’, that is, allow reverse conduction.

In IGBT circuits, a high-voltage parallel diode is necessary to allow reverse current flow. This is an extra cost and the diodes need to be high performance with minimal reverse recovery energy loss. SiC-MOSFETs, on the other hand, have an integral reverse diode but performance is relatively poor with a high forward voltage drop and significant recovery losses at operating temperatures. The SiC-FET, however, allows reverse conduction through the channel with no stored charge effects and low forward drop, effectively across the already low ON-resistance. The stacked Si-MOSFET in the package also conducts in reverse but being an optimised low voltage type, its body diode drop is small and also contributes little in recovery losses.

The latest generation SiC-FETs show lower losses than the traditional IGBT approach with additional side benefits. Table 2 shows calculated losses at six power levels comparing a current state-of-the-art IGBT module and parallel diode approach with SiC-FET versions.
page6image36304496

Table 2: Total conduction and switching losses comparing IGBTs and SiC-FETs – EV applications

SiC-FETs consistently achieve nearly a four-fold reduction in power loss at the typical 50-100kW level and nearly three-fold reduction at 200kW. In EV applications, this equates to more energy available for extended range and reduced cooling requirements leading to smaller and lighter heatsinking, leading in turn to reduced load on the vehicle and better range—a virtuous circle. The availability of these low resistance devices in low cost discrete packages allow very economical inverter construction.

Although wide band-gap technology is relatively new, there have been understandable concerns about practical reliability. Latest generation SiC-FET parts now have extensive testing data and use mature production processes to guarantee robustness. They also have built-in advantages; apart from the inherent high-temperature capability of silicon carbide, SiC-FETs have a self-limiting avalanche drain voltage characteristic with the channel
self-biasing to active mode with overvoltage, absorbing transient energy up to several joules.
Another SiC-FET benefit is that they are robust with short circuits. High current through the channel resistance produces a negative JFET gate bias, tending to switch the device OFF. Through self-heating, the positive temperature coefficient of the channel resistance then reduces the short circuit current further. This effect makes SiC-FETs easy to parallel with automatic current balancing, aided further by the relative insensitivity of the stacked MOSFET threshold voltage and reverse recovery characteristics to temperature changes.

EV charging

SiC-FETs are also ideal in fast charger applications where they offer peak efficiency in PFC front ends, and in the main DC-DC conversion stage—both typically using phase- shifted full bridge or LLC topologies. SiC diodes are already being used to implement output rectification in high voltage chargers due to their low drop and absence of reverse recovery losses. This is because synchronous rectification (SR) using Si-MOSFETs is complex at high voltage and can give no loss saving over diodes. However, using SiC-FETS with low RDS(ON) may prove more advantageous.
For example, at a 100A operating current with a 50% duty
cycle, a SiC diode will have conduction losses of nearly 100W, but the UF3SC065007K4S will have conduction losses of just 45W. In addition, SR opens the possibility of bi-directional power flow, allowing the EV battery to return power to the grid for utility load-levelling for example, with corresponding financial benefits.

Solid state circuit breakers are an important application in EVs because isolating batteries during servicing and fault conditions is mandatory. With their normally ON characteristic, JFETs find a natural home here.

Backward compatibility

As UnitedSiC SiC-FETs are available in TO-247 three- and four-lead packages, they can be a drop-in replacement for many IGBTs and Si-MOSFETs in motor drives. This gives a significant boost in efficiency with little change in the circuit, apart from perhaps gate drive resistors and small snubbers to tailor the switching edges. Gate drive voltage requirements are non-critical, typically 0-12V. Other benefits such as reducing existing snubbers for lower loss and removing parallel diodes in designs that were originally IGBT-based could be considered. With the latest generation of low RDS(ON) devices from UnitedSiC, SiC- FETs are paving the way for the EV drivetrain revolution.


Source: Automotiveworld

1594913258636.png
 
Last edited:

RonM

Well-known member
First Name
Ronald
Joined
May 16, 2020
Messages
62
Reaction score
96
Location
Palmetto Bay, Fl
Vehicles
2015 F150 CrewCab, future Cybertruck
Occupation
Retired
Country flag
Did anyone get the number of that bus???:oops::oops::oops:
 
OP
TruckElectric

TruckElectric

Well-known member
First Name
Bryan
Joined
Jun 16, 2020
Messages
445
Reaction score
545
Location
Texas
Vehicles
Dodge Ram diesel
Occupation
Retired
Country flag
  • Thread starter
  • Thread Starter
  • #10
Insights on the Automotive Power Electronics Global Market to 2026 - Players Include Continental, Rohm Semiconductor & Texas Instruments Among Others

Dublin, Aug. 20, 2020 (GLOBE NEWSWIRE) -- The "Global Automotive Power Electronics Market Analysis 2020" report has been added to ResearchAndMarkets.com's offering.

The Global Automotive Power Electronics market is expected to reach $5.64 billion by 2026 growing at a CAGR of 6.8% during 2019 to 2026. Power Electronics is the branch of electronics deals to control power fluctuation in components such as transistors, monitoring diodes, and others. It consists of the solid-state circuit device that transfers the current from source to the load and enhances the conservation of energy by managing the power.

Automotive power electronics are used to control engines of automobiles originated in the automotive electronics for proper controlling and conversion. It offers a wide range of application such as central locking, anti-braking system, power steering, braking system, and seat control. Further, its application is extended to various fields such as aerospace, automotive, commercial, industrial, telecommunication, transportation, and utility systems. It allows the flow of the current in a bidirectional way, and thereby improves the efficiency of the system by withstanding high power temperature.

The factors such as increasing demand for vehicle connectivity, infotainment and powertrain electrification, increase in demand for energy-efficient battery-powered devices, increasing adoption of safety systems, and growing environmental concerns are driving the market growth. However, increase in the overall cost of the vehicle is expected to restrain the market growth during the forecast period.

Depending on vehicle type, the passenger vehicle segment makes a major contribution to the automotive power electronics market. The increased demand for safety systems and fuel efficient technology associated with vehicle power train in emerging markets are fueling factor for the dominance of passenger vehicle. From past 5 or 10 years, passenger vehicles are getting installed with number of safety, comfort, entertainment and vehicle management features. This is because of factor such as changing preferences of vehicle buyer supported by government legislations. Moreover, increasing population in developing economies also encourage OEMs to add specific features in new models of passenger vehicles.


The key vendors mentioned are Continental, Rohm Semiconductor, Texas Instruments, Robert Bosch, Toshiba Corp, ON Semiconductor, Infineon Technologies, Mitsubishi Electric, NXP Semiconductors, Qualcomm, ACTIA Group, STMicroelectronics, Renesas Electronics Corp, Vishay Intertechnology, Fuji Electric, BYD, Delphi, Delta Electronics, Denso, and Semikron.

Components Covered:

  • Sensors
  • Microcontroller Unit (MCU)
  • Power Integrated Circuit
Devices Covered:

  • Power Module
  • Power IC
  • Power Discrete
Materials Covered:

  • Silicon Carbide
  • Gallium Nitride
  • Silicon
  • Other Materials
Vehicle Types Covered:

  • Commercial Vehicle
  • Passenger Vehicle
  • Electric Vehicle
Electric Vehicle Types Covered:

  • Battery Electric Vehicles (BEV)
  • Plug-in Hybrid Electric Vehicle (PHEV)
  • Hybrid Electric Vehicle (HEV)
  • IC Engine Vehicle
  • Pure Electric Vehicle
  • Fuel Cell Electric Vehicle (FCEV)
Applications Covered:

  • Engine Management & Powertrain
  • Battery Management
  • Advanced Driver Assistance System (ADAS)
  • Body Control & Comfort
  • Telematics
  • Infotainment
  • Powertrain & Chassis
  • Safety and Security Electronics
  • Body Electronics
  • Engine Electronic
  • Transmission Electronics
  • Active Safety
  • Driver Assistance
  • Passenger Comfort
  • Entertainment System
Source: RESEARCH AND MARKETS

Key players in the Next-Generation Power Semiconductors covers :
Check Point Software Technologies
Sophos Group
Watchguard Technologies
Juniper Networks
Fuji Electric
NXP Semiconductors
Fairchild
Zscaler
Cisco Systems
Barracuda Networks
Renesas Electronics
Toshiba
STMicroelectronics
Palo Alto Networks
Semikron
Forcepoint
Infineon Technologies
Mitsubishi Motors
Fortinet
Vishay Intertechnology

Analysts have also stated the research and development activities of these companies and provided complete information about their existing products and services. Additionally, the report offers a superior view over different factors driving or constraining the development of the market.

The Next-Generation Power Semiconductors can be split based on product types, major applications, and important countries as follows:

The basis of applications, the Next-Generation Power Semiconductors from 2015 to 2025 covers:

Renewable Energy
Hybrid & Electric Vehicle
Smart Homes
LED Lights

The basis of types, the Next-Generation Power Semiconductors from 2015 to 2025 is primarily split into:
GaN
SiC

Source: startupng
 
Last edited:
OP
TruckElectric

TruckElectric

Well-known member
First Name
Bryan
Joined
Jun 16, 2020
Messages
445
Reaction score
545
Location
Texas
Vehicles
Dodge Ram diesel
Occupation
Retired
Country flag
  • Thread starter
  • Thread Starter
  • #11
ROHM’s New 4th Generation SiC MOSFETs Featuring the Industry’s Lowest ON Resistance
Advanced design expected to see widespread adoption in the main drive inverters of EVs


883898.jpg

ROHM's 4th Generation 1200V SiC MOSFETs optimized for automotive powertrain systems, including the main drive inverter, as well as power supplies for industrial equipment.


Kyoto, Japan and Santa Clara, CA, June 17, 2020 (GLOBE NEWSWIRE) -- ROHM announces the cutting-edge 4th Generation 1200V SiC MOSFETs optimized for automotive powertrain systems, including the main drive inverter, as well as power supplies for industrial equipment.

In recent years, the proliferation of next-generation electric vehicles (xEVs) has been accelerating the development of smaller, lighter, and more efficient electrical systems. In particular, improving efficiency while decreasing the size of the main inverter that plays a central role in the drive system remains among the most important challenges, requiring further advancements in power devices.

The capacity of the onboard battery is increasing to improve the cruising range of EVs. And in conjunction with this, the use of higher voltage batteries (800V) is progressing to meet the demand for shorter charging times.

To solve these various challenges, designers urgently need SiC power devices capable of providing high withstand voltage with low losses. ROHM, a pioneer in SiC, began mass producing SiC MOSFETs ahead  of the industry in 2010. From early on, ROHM has strengthened its considerable lineup to include AEC-Q101 qualified products allowing the company to hold a large market share for automotive onboard chargers (OBC).

For power semiconductors there is often a trade-off relationship between lower ON resistance and short-circuit withstand time, which is required to strike a balance for achieving lower power losses in SiC MOSFETs. ROHM was able to successfully improve this trade-off relationship and reduce ON resistance per unit area by 40% over conventional products without sacrificing short-circuit withstand time by further improving an original double trench structure. In addition, significantly reducing the parasitic capacitance (which is a problem during switching) makes it possible to achieve 50% lower switching loss over our previous generation of SiC MOSFETs.

As a result, ROHM’s new 4th Generation SiC MOSFETs are capable of delivering low ON resistance with high-speed switching performance, contributing to greater miniaturization and lower power consumption in a variety of applications, including automotive traction inverters and switching power supplies. Bare chip samples have been made available from June 2020, with discrete packages to be offered in the future.

ROHM is committed to continue to expand its SiC power device lineup while combining modularization technologies with peripheral devices such as control ICs designed to maximize performance in order contribute to technical innovation in next-generation vehicles. At the same time, ROHM will provide solutions that resolve customer issues – including web-based simulation tools that reduce application development man-hours and help prevent evaluation problems.

Key Features

1) Improved trench structure delivers the industry’s lowest ON resistance


In 2015, ROHM began mass production of the industry-first trench-type SiC MOSFETs utilizing an original structure. Now, ROHM has successfully reduced ON resistance by 40% compared to conventional products without sacrificing short-circuit withstand time by further improving its original double trench structure.

2) Achieves lower switching loss by significantly reducing parasitic capacitance


Generally, lower ON resistances and larger currents tend to increase the various parasitic capacitances in MOSFETs, which can inhibit the inherent high-speed switching characteristics of SiC.

However, ROHM was able to achieve 50% lower switching loss over conventional products by significantly reducing the gate-drain capacitance (Cgd).


Terminology

MOSFETs (Metal Oxide Semiconductor Field Effect Transistors)


The most commonly used structure in FETs. Often adopted as switching elements.

Short-Circuit Withstand Time

Indicates the time it takes for a MOSFET to fail due to a short-circuit. Normally, when a short-circuit occurs, a large current exceeding the maximum rating will flow, which can lead to abnormal heat generation, thermal runaway and ultimately, destruction. Longer short-circuit withstand time is in a trade-off relationship with higher performance characteristics, such as ON resistance.

Double Trench Structure

ROHM’s original trench structure. Although the adoption of a trench structure to SiC MOSFETs was shown to be effective in reducing ON resistance, it was necessary to mitigate the electric field generated in the trench gate section to ensure long-term reliability of device. In response, ROHM adopted a unique double-trench structure that minimizes electric field concentration, allowing it to become the first supplier to mass produce trench-type SiC MOSFETs in 2015.

Parasitic Capacitance

Inherent capacitance that occurs due to the physical structure within electronic components. In the case of a MOSFET, there is a gate-source capacitance (Cgs), gate-drain capacitance (Cgd), and drain-source capacitance (Cds). Cgs and Cgd are determined by the capacitance of the gate oxide film, while Cds is the junction capacitance of the parasitic diode.

Trench Structure

The word ‘trench’ means a narrow excavation or groove. This design involves forming a groove on the chip surface and the gate on the MOSFET side wall. JFET resistances don’t exist compared to a planar-type MOSFET configuration, making it possible to achieve a finer structure over planar topologies – resulting in an ON resistance close to the original performance of the SiC material.

Source: Globenewswire
 
Last edited:

ldjessee

Well-known member
First Name
Lloyd
Joined
Apr 22, 2020
Messages
424
Reaction score
306
Location
Indiana, USA
Vehicles
reservation for 2 motor Cybertruck, Nissan Leaf, Subau Outback, Kawasaki Vulcan 1700 Vaquero ABS
Occupation
programmer
Country flag
OP
TruckElectric

TruckElectric

Well-known member
First Name
Bryan
Joined
Jun 16, 2020
Messages
445
Reaction score
545
Location
Texas
Vehicles
Dodge Ram diesel
Occupation
Retired
Country flag
  • Thread starter
  • Thread Starter
  • #15
@TruckElectric

You need to tl;dr this stuff man.
P7MjC7Zcz_vAf-mZmBiaWKotwa4LZJIvRx608FJ5TwjHHYecEg.gif

Building Power Electronics With Microscopic Plumbing Could Save Enormous Amounts of Money

Designing semiconductor circuits hand-in-hand with microfluidic cooling systems could mean huge boosts in efficiency

MzY4NjQ2MQ.jpg
Illustration: Vytautas Navikas/EPFLLiquid coolant flows from fins on the back of a silicon substrate (gray) into tiny microchannels in the substrate. Researchers carve the microchannels via tiny slits etched from the top of the gallium nitride (gold) layer that contains electronic devices

The heat generated by today’s densely-packed electronics is a costly resource drain. To keep systems at the right temperature for optimal computational performance, data center cooling in the United States consumes the as much energy and water as all the residents of the city of Philadelphia. Now, by integrating liquid cooling channels directly into semiconductor chips, researchers hope to reduce that drain at least in power electronics devices, making them smaller, cheaper and less energy-intensive.

Traditionally, the electronics and the heat management system are designed and made separately, says Elison Matioli, an electrical engineering professor at École Polytechnique Fédérale de Lausanne in Switzerland. That introduces a fundamental obstacle to improving cooling efficiency since heat has to propagate relatively long distances through multiple materials for removal. In today’s processors, for instance, thermal materials syphon heat away from the chip to a bulky, air-cooled copper heat sink.

For a more energy-efficient solution, Matioli and his colleagues have developed a low-cost process to put a 3D network of microfluidic cooling channels directly into a semiconductor chip. Liquids remove heat better than air, and the idea is to put coolant micrometers away from chip hot spots.

But unlike previously reported microfluidic cooling techniques, he says, “we design the electronics and the cooling together from the beginning.” So the microchannels are right underneath the active region of each transistor device, where it heats up the most, which increases cooling performance by a factor of 50. They reported their co-design concept in the journal Naturetoday.

Researchers first proposed microchannel cooling back in 1981, and startups such as Cooligy have pursued the idea for processors. But the semiconductor industry is moving from planar devices to 3D ones and towards future chips with stacked multi-layer architectures, which makes cooling channels impractical. “This type of embedded cooling solution is not meant for modern processors and chips, like the CPU,” says Tiwei Wei, who studies electronic cooling solutions at Interuniversity Microelectronics Centre and KU Leuven in Belgium. Instead, this cooling technology makes the most sense for power electronics, he says.

Power electronics circuits manage and convert electrical energy, and are used widely in computers, data centers, solar panels, and electric vehicles, among other things. They use large-area discrete devices made from wide-bandgap semiconductors like gallium nitride. The power density of these devices has gone up dramatically over the years, which means they have to be “hooked to a massive heat sink,” Matioli says.

More recently, power electronics modules have turned to liquid cooling either via cold plates or micro channel cooling systems. But all microchannel cooling systems so far have been made separately and then bonded to the chip. The bonding layer adds heat resistance, and the channels and circuit devices aren’t aligned closely.

“We take it to the next level,” says Matioli, by making the device and cooling channels in the same chip. They etch micrometer-wide slits in a gallium nitride layer coated on a silicon substrate. The slits are 30µm-long and 115µm-deep. Using a special gas etching technique, they widen the slits in the silicon substrate to form the channels through which liquid coolant is pumped.

Then the researchers seal the tiny openings in the gallium nitride layer with copper, and make devices on top. “We only have microchannels on the tiny region of wafer that’s in contact with each transistor,” he says. “That makes the technique efficient because we can extract a lot of heat due to proximity but we use very little pumping power.”

As a demonstration, the researchers made an AC-to-DC rectifier circuit composed of four Schottky diodes, each capable of handling 1.2kV. A circuit like this would typically require a fist-sized heat sink and. But with the integrated liquid cooling system, the circuit chip sits on a USB stick-sized printed circuit board, which is made of three layers and has channels carved into it to deliver coolant to the chip.

They show that hot spots with power densities of over 1,700 watts per square centimeter can be cooled using only 0.57 W/cm2 of pumping power. That’s a 50-fold increase in performance compared to previously reported microfluidic channel cooling.

The reliability of the gallium nitride membranes and copper sealed layer should be investigated over time, Wei says. But this innovative cooling solution is a big step towards “low-cost, ultra-compact and energy-efficient cooling systems for power electronics.”

Source: IEEE SPECTRUM
 

Advertisement




Advertisement







 
Top