Power unit with local filtration and multi-component converter with local filtration
Patent Information
- Application Number
- DE602022008172
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2021-12-28
- Filing Date
- 2022-12-28
- Publication Date
- 2024-11-27
- Estimated Expiration
- 2042-12-28
AI Technical Summary
Current solutions for reducing common mode disturbances in power electronics are bulky and heavy due to the use of inductive filters, and the parasitic capacitances in electronic switches exacerbate these issues, especially with increasing switching frequencies.
A component with a conductive plane isolated from the sole and substrate, connected to a reference voltage via a specific impedance or inductance, is used to filter currents through parasitic capacitances, reducing leakage currents by damping parasitic resonance frequencies.
This configuration effectively attenuates parasitic currents by shifting the resonance frequency and dissipating stray current energy, reducing the amplitude of undesirable frequencies and improving thermal management without significant weight or volume increase.
Description
Background to the invention
[0001] In power electronics, many static converters have been developed, mainly inverters and rectifiers. More complex equipment such as switching power supplies are also available. These converters are equipped with electronic components that allow switching an electrical signal. Switching components are often referred to as "electronic switches." These components include diodes, transistors, thyristors, triacs, etc.
[0002] Power diodes were developed from the 1950s. In the 1960s, thyristors and power transistors appeared. Later, in the 1980s, insulated gate bipolar transistors (IGBTs) appeared. This type of transistor combines the advantage of control by a field effect transistor with the low conduction losses of a bipolar transistor. To date, IGBTs can switch powers of around 10 MW at speeds above 10 kHz. For lower powers, MOSFET transistors are also widely used. In the 2000s, components, and in particular MOSFET transistors, based on silicon carbide (SiC) appeared. They can achieve speeds above 1 MHz.More recently, gallium nitride GaN has also been introduced into MOSFET transistors to make electronic switches. Gallium nitride allows for even higher switching speeds at frequencies above 10 MHz.
[0003] By switching electrical power signals, these components generate heat losses that must be dissipated. Electronic switches are most often cooled using radiators that dissipate heat by convection of air flowing along the radiator walls. Radiators can also be equipped with channels through which a refrigerant circulates.
[0004] To ensure good heat transfer from the component to its heat sink, it is necessary to minimize the thermal resistance of the path taken by the heat from its emission zone in the core of the component to the heat sink allowing the dissipation of this heat. The electronic switch most often has a metal base plate placed in the immediate vicinity of the core of the component and through which the heat to be evacuated passes. The base plate can be electrically insulated from the component or form one of its connection points. When mounting the electronic switch, the base plate is pressed against the heat sink. The contact of the base plate on the heat sink can be thermally improved by means of a thermally conductive paste. It is also possible to interpose an electrical insulator between the base plate and the heat sink in the form of a plate based on aluminum oxide, mica or silicone.
[0005] The safety of users of equipment using electronic switches generally requires electrical grounding of radiators. For better heat dissipation, it is even possible to use the outer casing of the equipment as a radiator. This further requires connecting the casing to the electrical ground of the system containing the equipment.
[0006] There figure 1represents an electrical architecture 10 comprising two static converters: a rectifier 12 and an inverter 14. The electrical architecture 10 is well suited to equipping a vehicle, in particular an aircraft. The electrical architecture 10 comprises an AC network 16 making it possible to supply a load 18, such as for example a motor through the rectifier 12, generating a DC voltage from the AC voltage from the AC network 16, and through the inverter 14 providing an AC voltage to the load 18 from the DC voltage.
[0007] On the figure 1, the electrical architecture 10 is represented in a summary manner. In practice, the direct voltage can form a network making it possible to power a large number of loads, each of which can operate with a different alternating voltage. It is possible to adapt the frequency of the alternating voltage to the needs of the load. For example, the speed of the motor can be modulated by varying the frequency using the inverter 14. Several inverters can power the same load. The electrical architecture 10 can be reconfigurable in real time. It makes it possible to power a load using as many converters as necessary, as for example described in the patent application published under No. EP 2 887 519 and filed in the name of the applicant. Document US 2020 / 0161207 A1 describes adaptations for a semiconductor power module making it possible to reduce parasitic capacitances and common mode interference.
[0008] Common mode disturbances can circulate in the electrical architecture 10. These disturbances represented in thick dotted lines on the figure 1 circulate mainly through an electrical ground 20 of the electrical architecture 10. The alternating network 16 being connected to the ground 20, the disturbances are superimposed on the alternating voltages of the alternating network 16 then on the direct voltages coming from the rectifier 12.
[0009] To date, common mode disturbances are limited by means of inductive filters placed, for example, on the DC network. As illustrated in the figure 1, a common mode filter 22 comprises two inductors arranged on the same magnetic core. In each of the inductors circulates a phase of the DC network. The inductors are made by means of electrical conductors wound on the common magnetic core. The electrical conductors forming the inductors are sized to support the useful current of the DC network. As a result, the conductor sections are large. The magnetic core on which the conductors are wound is also large. This results in a significant weight and volume for the filter 22.
[0010] Furthermore, the electronic switches of the converters have a parasitic capacitance 24 formed between the conduction channel of the switch, the base and the radiator. This capacitance 24 is due to the proximity between the conduction channel, the base and the radiator, proximity made necessary for the evacuation of the heat released during switching.
[0011] Internal experiments have shown that common mode disturbances tend to propagate through the parasitic capacitances 24 of the electronic switches to reach the electrical ground 20. On the figure 1, only the parasitic capacitances 24 of the electronic switches of the inverter 14 are represented. The diodes of the rectifier 12 can also have the same type of parasitic capacitance, as soon as one of their electrodes is arranged near or in contact with a mechanical part electrically connected to the electrical ground 20. The currents forming the common mode disturbances are all the more important as the switching frequency increases. With the implementation of electronic switches capable of achieving extremely reduced switching times, disturbance currents are becoming increasingly worrying. Due to their volume and mass, current solutions for reducing disturbances, in particular common mode currents, are difficult to implement with new generations of electronic switches. Summary of the invention
[0012] The invention proposes a solution to these problems by filtering at the source the currents which can pass through the parasitic capacitances of electronic switches.
[0013] To this end, the invention relates to a component configured to switch an electrical signal, the component comprising: an insulating substrate carrying a semiconductor chip ensuring signal switching, a base plate on which the substrate is fixed, the base plate being configured to dissipate heat emitted during switching of the component, a conductive plane arranged between the base plate and the insulating substrate, the conductive plane being electrically insulated from the base plate, a specific component having an impedance of at least 1 Ohm and / or an inductance of at least 1 µH, via which the conductive plane is connected to a reference voltage.
[0014] Advantageously, the component is a resistor of at least 1 Ohm.
[0015] The invention also relates to a static converter comprising several components according to one of the preceding claims in which the conductive planes and the bases of each component are common.
[0016] Advantageously, the converter comprises a first group and a second group of components, in each of the groups of components, the conductive planes and the soles of each component are common, the conductive plane of the first group is connected to a first reference voltage via a first inductance and the conductive plane of the second group is connected to a second reference voltage via a second inductance, and the two inductances are coupled and have an impedance of at least 1 µH. Brief description of the figures
[0017] The invention will be better understood and other advantages will appear on reading the detailed description of an embodiment given by way of example, a description illustrated by the attached drawing in which: there figure 1 , already described, represents an electrical architecture in which the invention can be implemented; the figure 2 represents a switching component according to the invention; the figure 3 represents an equivalent diagram of the component of the figure 1 ; there figure 4 represents several switching components of a static converter according to the invention; the Figure 5 schematically represents a static converter according to the invention.
[0018] For the sake of clarity, the same elements will have the same references in the different figures. Detailed description of the invention
[0019] There figure 2represents in section a switching component 30 implementing an active chip 32 based on semiconductor. Any type of switching component can be implemented within the scope of the invention. The component can be naturally switched such as a diode or controlled switching such as a thyristor or a transistor. To date, many materials are used to produce switching components, in particular silicon, germanium, gallium arsenide, silicon carbide, gallium nitride. The invention can be implemented regardless of the semiconductor material used.
[0020] There figure 2schematically represents in section the component 30. The chip 32 is deposited on an insulating substrate 34, for example ceramic. A metallization selectively covers the substrate 34 and allows the chip 32 to be electrically connected. The metallization here comprises two parts 36 and 38. The part 36 receives a pin 40 and the part 38, a pin 42. The part 36 extends under the chip 32 and forms a first electrode of the component 30. The part 38 is electrically connected to the upper part of the chip 32 by means of one or more metal legs 44 forming a second electrode of the component 30. An electrical signal can circulate between the two pins 40 and 42 depending on the state of the junction in the semiconductor forming the chip 32. In other words, the component 30 ensures the switching of the electrical signal.For a naturally switching component, the potential difference between the two pins 40 and 42 determines the state of the junction and for a controlled switching component, a control electrode, not shown in the . figure 2 , allows the switching to be triggered by switching the junction from a blocked state to a conducting state. The control electrode can be connected to another part of the metallization by means of another leg not located in the cutting plane of the figure 2 .
[0021] During its operation, the component 30 releases heat forming losses that must be evacuated. This heat is mainly released during switching and when the junction is conducting due to its internal resistance. The heat generated in the chip 32 is mainly drained towards a base plate 46, for example made of a metal alloy for its good capacity to conduct heat. The base plate 46 can for example be made of a copper or aluminum alloy. The base plate 46 can be fixed to a radiator, not shown in the figure 1and allowing the heat to be evacuated. The evacuation can be done into the ambient air by natural or forced convection. The evacuation can also be done by means of one of the channels circulating in the radiator and in which a heat transfer fluid circulates. The sizing of the radiator is carried out according to the heat generated by the component 30 during its use and, for evacuation by convection, according to the ambient environment in which the radiator is immersed. It is common to connect the radiator to the electrical ground of the equipment in which the component 30 is implemented, for example a converter 12 or 14 as shown in the figure 1 .
[0022] As mentioned above, a parasitic capacitance is formed between, on the one hand, the chip 32 and the parts 36 and 38 of the metallization, and on the other hand, the base 46, the substrate 34 forming a dielectric of this parasitic capacitance. Another parasitic capacitance is formed in the chip 32 itself between its electrodes. A displacement current can be formed between the junction and the base 46 through the parasitic capacitance. This current can be all the more important as the switching speed increases. In practice, during switching, the current flowing in the junction is established and interrupted by forming a current step in a timing diagram. The applicant has noticed that the repetition of these current steps generates significant leakage currents flowing in the electrical mass of the system through the parasitic capacitance of the component 30.The invention seeks to limit the current flowing in this parasitic capacitance by means of a filter placed as close as possible to the chip 32.
[0023] According to the invention, a conductive plane 50 is arranged between the sole 46 and the insulating substrate 34. The conductive plane 50 is electrically insulated on the one hand from the chip 32 and the two parts 36 and 38 of the metallization and on the other hand from the sole 46. On the chip and metallization side, the insulation of the conductive plane 50 is ensured by the substrate 34. On the sole 46 side, an insulating film 52 can be interposed between the conductive plane 50 and the sole 46. The conductive plane 50 can be a metallic film, for example made of a copper or aluminum alloy, deposited on the face of the substrate 34 opposite that receiving the metallization. The conductive plane 50 can be made of any other type of conductive material, for example carbon-based. The conductive plane 50 can be continuous or made in the form of a grid having openings. The section of the openings is defined as a function of the wavelength of the parasitic signals that it is desired to filter.The insulating film 52 is of course chosen to ensure the electrical insulation of the conductive plane 50 with respect to the sole 46. It is advantageous to choose a material having good thermal conduction properties to avoid slowing down the thermal transfer of the heat energy dissipated by the chip 32 to the sole 46. For this purpose, an insulating film 52 based on silicone or mica can be chosen. The electrical insulation of the conductive plane 50 with respect to the sole 46 can also be ensured by means of a layer, for example of the varnish type, deposited on the conductive plane 50 opposite the sole 46. The level of electrical insulation between the conductive plane 50 and the sole 46 is defined as a function of the potential difference that may exist between the conductive plane 50 and the sole 46. This potential difference will be specified later.
[0024] There figure 3represents in the form of an electronic diagram the switching component and the parasitic capacitances connecting it to the base 46. The component 30 is represented diagrammatically by a transistor. As mentioned above, other switching components can be implemented within the framework of the invention. A first capacitance 54 is formed between, on the one hand, the chip 32 and its metallization and, on the other hand, the conductive plane 50. The substrate 34 forms the dielectric of the capacitance 54. A second capacitance 56 is formed between the conductive plane 50 and the base 46 which is connected to the potential of a reference voltage such as an electrical ground. The insulating film 52 forms the dielectric of the capacitance 56. Given the arrangement of the conductive plane 50 between the substrate 34 and the base 46, the two capacitances 54 and 56 can be considered as connected in series.
[0025] The component 30 further comprises an electrical contact 58 configured to electrically connect the conductive plane 50. The connection makes it possible to produce a damping filter with an electronic component which may be external to the component 30. In the diagram of the figure 3, the electrical contact 58 appears at the common point of the two capacitors 54 and 56. For example, the conductive plane 50 can be connected to an electrical ground connection of the component 30, via a resistor R. The electrical ground connection 60 can be connected to any reference voltage. By connecting the electrical ground connection 60 to the potential of the base plate 46, there is little potential difference across the terminals of the capacitor 56. It is thus possible to reduce the thickness of the insulation between the conductive plane 50 and the base plate 46. The conductive plane 50 could be connected directly to a reference voltage. The parasitic capacitors 54, 56 and the connection wires already have resistances whose value is typically of the order of a few milliohms. Coupled with the capacity added by the presence of the conductive plane, this low resistance allows at most to shift the parasitic resonance frequency due to the parasitic capacitors of the component.Damping is virtually non-existent. During internal tests, the applicant found that a resistance of at least one Ohm gave good results in attenuating the amplitude of certain unwanted frequencies. More specifically, in an inverter without the conductive plane, stray currents circulating in the ground and presenting amplitude peaks around 1 MHz were observed. By implementing a conductive plane associated with a resistance of one Ohm for each switching component of the inverter, the peaks were attenuated by around twenty dB.
[0026] Furthermore, by implementing a specific component to produce the resistance R, it is possible to arrange it at a chosen location, so that the dissipation of the energy of the parasitic current can be done for example on a heat sink provided for this purpose. The resistance value of the specific resistive component R of at least one Ohm is added to the resistances of the parasitic capacitances and the connection wires of the conductive plane 50 in its connection to the reference voltage. As mentioned above, the resistance value of the specific resistive component R is of the order of 1000 times greater than that of the resistances of the parasitic capacitances and the connection wires. Consequently, almost all of the energy of the parasitic current is dissipated in the specific component. Other types of passive or even active components can be connected to the electrical contact 58.In particular, passive components of the capacitance and / or inductance type can be connected between the electrical contact 58 and a reference voltage in addition to or instead of the resistance R. In the case of an inductance, during tests carried out internally, the applicant found that a value of at least 1 µH gave good results in participating in the damping of the parasitic current. In the case of a wound inductance, it is found that its value tends to decrease when the current frequency increases mainly due to the skin effect. The inductance value of 1 µH beyond which the results are sensitive is measured at 10 kHz.
[0027] It is also possible to combine a specific resistive component R and a specific inductive component L with the capacitance 56 to obtain a damping circuit allowing both to shift the resonant frequency and to dampen the harmonics of the parasitic current.
[0028] These components can be arranged in the housing of the switching component 30. The electrical contact is then an extension of the conductive plane projecting from the substrate 34. The electrical contact 58 can also be a part of the conductive plane 50. The resistor R or, more generally, the component connected to the electrical contact 58 can be surface-mounted on the substrate 34. Alternatively, the electrical contact 58 can come out of the component and allow a designer of equipment in which one can for example find the rectifier 12, the converter 14 and the load 18 to choose a component to be connected to the electrical contact 58, a component adapted to the equipment and to the parasitic currents that he observes.
[0029] There figure 4 represents several switching components having a common sole 46. The Figure 5represents in the form of an electronic diagram a converter 14 of the three-phase inverter type in which the invention is implemented. As indicated above, the invention can be implemented in any type of converter, regardless of the number of switching components. The Figures 4 and 5 are interested in converters implementing several switching components. In these converters, it is advantageous to share the conductive plane which becomes common to the different switching components or at least to a group of several switching components of the converter. The Figure 5 describes an inverter. It is understood that the sharing of a conductor plane can be applied to any type of converter, for example in a rectifier.
[0030] There figure 4 represents two switching components 30. Unlike component 30 described in isolation using the figure 2 , on the figure 4, the sole 46 and the conductive plane 50 are common to the two switching components 30. On the figure 4 , the insulating film 52 is also common to the two switching components 30. As previously, any other means of insulating the conductive plane 50 from the sole 46 remains possible. The electrical contact 58 is associated with the conductive plane 50 and is also common to the two switching components 30.
[0031] The inverter 14, shown schematically on the Figure 5, makes it possible to supply three phases 14u, 14v and 14w in alternating voltage from two direct voltages 14+ and 14-. The inverter 14 comprises two groups of switching components. The first group comprises transistors marked T+ and making it possible to switch the direct voltage 14+. Each transistor T+ of the first group contributes to the supply of one of the phases 14u, 14v and 14w. Similarly, the second group comprises transistors marked T- and making it possible to switch the direct voltage 14-. Each transistor T- of the second group contributes to the supply of one of the phases 14u, 14v and 14w.
[0032] The three transistors T+ of the first group have a common conductive plane 50+ and the three transistors T- of the second group have a common conductive plane 50-. The converter 14 may comprise only one sole common to all the transistors T+ and T-. The sole is for example connected to an electrical ground of an equipment in which the converter 14 is integrated. Alternatively, it is possible to provide two separate soles 46+ and 46-, one per group of transistors T+ on the one hand and T- on the other hand. The two soles 46+ and 46- may be connected to the same reference voltage or to two separate reference voltages. For example, the sole 46+ associated with the transistors T+ may be connected to the DC voltage 14+ and the sole 46- associated with the transistors T- may be connected to the DC voltage 14- as illustrated in the Figure 5It is also possible to connect the 46+ sole to the 14- DC voltage and the 46- sole to the 14+ DC voltage.
[0033] The presence of two conductive planes 50+ and 50- results in the presence of two capacitors, respectively 54+ and 54-, on the one hand between the transistors T+ and the conductive plane 50+ and on the other hand between the transistors T- and the conductive plane 50-. Associated with each conductive plane 50+ and 50-, there is an electrical contact 58, respectively 58+ and 58-. On the Figure 5 , we find two capacities 56+ and 56- formed between the conductive plane, respectively 50+ and 50- and their corresponding base or the base common to all transistors T+ and T-.
[0034] It is easy to transpose the described scheme from the Figure 5to other types of converter, for example to the full bridge rectifier 12, i.e. comprising four rectifier diodes. It is possible to define a first group of two diodes switching one of the phases of an alternating voltage and a second group of two diodes switching the other phase of the alternating voltage. In such a rectifier, a first conductive plane may be common to the first group of diodes and a second conductive plane may be common to the second group of diodes.
[0035] We saw, with the help of the figure 3, that the conductive plane could be connected to a reference voltage via any type of component, active or passive. The reference voltage can be, for example, the electrical ground 60 of the equipment in which the inverter 14 is implemented. The reference voltage can also be one of the DC voltages 14+ or 14-. It is also possible to dissociate the reference voltage. More precisely, the electrical contact 58+ can be connected, directly or via components of type R, L, C, to the DC voltage 14+. Furthermore, the electrical contact 58- can be connected, directly or via components of type R, L, C, to the DC voltage 14-. In the inverter 14, the conductive plane 50+ is connected to a reference voltage via an inductance L+ and the conductive plane 50- is connected to the same reference voltage 60 via an inductance L-. On the Figure 5, the two inductances L+ and L- have a common point connected to the same reference voltage. Alternatively, it is possible to dissociate the reference voltages to which the two inductances L+ and L- are connected. For example, the inductance L+ can be connected to the DC voltage 14+ and the inductance L- can be connected to the DC voltage 14-.
[0036] The two inductances L+ and L- can be coupled, for example, by winding them on the same magnetic core. In addition to the two inductances L+ and L-, it is possible to add other passive or active components. As mentioned above, it is possible to connect passive components such as resistance and / or capacitance and / or inductance between the common point of the two inductances, or between each inductance and the respective reference voltage.
[0037] The implementation of two inductances coupled in common mode can be carried out for any other type of converter and in particular for the rectifier (12).
Claims
1. A component which is configured to switch an electrical signal, the component comprising: - an insulating substrate (34) bearing a semiconductor chip (32) which ensures switching of the signal; - a sole plate (46) on which the substrate (34) is secured, the sole plate (46) being configured to discharge heat emitted during switching of the component (30; T+, T-); - a conductive plane (50; 50+, 50-) positioned between the sole plate (46) and the insulating substrate (34), the conductive plane (50; 50+, 50-) being electrically insulated against the sole plate (46), characterised in that it further comprises: - a specific component (R, L+, L-) with an impedance of at least 1 Ohm and / or an inductance of at least 1 µH, by means of which the conductive plane (50; 50+, 50-) is connected to a reference voltage (14+, 14-; 60).
2. The component according to claim 1, wherein the specific component is a resistor (R) of at least 1 Ohm.
3. A static converter comprising multiple components (T+, T-) according to any one of the preceding claims, wherein the conductive planes (50+, 50-) and the sole plates (46) of each component (T+, T-) are shared.
4. The converter according to claim 3, comprising a first group (T+) and a second group (T-) of components according to any one of claims 1 or 2, wherein, in each of the groups of components, the conductive planes (50+, 50-) and the sole plates (46+, 46-) of each component (T+, T-) are shared, wherein the conductive plane (50+) of the first group (T+) is connected to a first reference voltage (14+) by means of a first inductor (L+), and the conductive plane (50-) of the second group (T-) is connected to a second reference voltage (14-) by means of a second inductor (L-), and wherein the two inductors (L+, L-) are coupled, and have an impedance of at least 1 µH.
