Transformer-driven switching device and transformer-driven power switch system

Abstract

Transformer-driven switching devices and transformer-driven power switch systems are disclosed, where the transformer-driven switching devices are used to switch high currents. These devices include power switches, such as gallium nitride (GaN) transistors. Transformers are used to transfer both control timing and power used to control the power switches. These transformers can be coreless, such that they can be integrated within a silicon die. Rectifiers, pull-down control circuits, and their related components are preferably integrated in the same die as the power switches, for example, in a GaN die, such that the transformer-driven switching devices will be fully included in silicon and GaN dies, without necessarily requiring a cored transformer, an auxiliary power supply, or a level shifting circuit.

Description

Technical Field

[0001] This application relates to circuit topologies and related apparatus for driving power switches, wherein transformers and rectifiers provide both control timing and power for the control terminals of the power switches. Background Technology

[0002] Many modern electronic circuits, such as switched-capacitor converters and motor drivers, utilize power switches to control the flow of high current levels. Power switches typically require driver circuitry for supplying and drawing current to and from the control terminals of the power switch. Drivers for power switches with load terminals (e.g., sources) directly coupled to ground may be relatively simple, as the circuitry of such drivers may be ground-referenced. However, power switches are typically not directly coupled to ground. For example, a half-bridge configuration may include a low-side power switch coupled between ground and the switching node, and a high-side power switch coupled between the positive voltage rail and the switching node. In other examples, switched-capacitor converters or multi-stage Class-D amplifiers may include a series of cascaded power switches between ground and the positive voltage rail, where the voltage at the load terminals (e.g., sources) of intermediate power switches typically falls between ground and the positive voltage rail.

[0003] Power switches not coupled to ground, such as high-side power switches or other floating power switches, present unique challenges for driving. Consider, for example, an n-channel metal-oxide-semiconductor field-effect transistor (MOSFET) power switch with source, drain, and gate (control terminals). The turn-on voltage driven to the MOSFET gate is referenced to the source voltage, which can be floating. The driver typically has a reference terminal for coupling to the MOSFET source and must be equipped with a power supply that provides power at a positive voltage, such as 5V, higher than the reference (source) voltage. Therefore, a floating power supply referenced to the MOSFET source voltage is required. Furthermore, driver inputs for digital signals, typically alternating between, for example, ground and 3V, generally require level shifting based on the MOSFET source voltage.

[0004] A typical technique for powering a high-side MOSFET driver uses a bootstrap power supply, which includes a diode and a relatively large capacitor. When the high-side MOSFET is off and its source is grounded, such as during the intervals when the low-side MOSFET is on, the capacitor charges from a low-voltage (ground-referenced) power supply via the diode. Once the high-side MOSFET source is disconnected from ground, such as when the low-side MOSFET is off, the capacitor powers the high-side MOSFET driver as needed to turn it on. Another technique uses a transformer, where a core is used to transfer power from a ground-referenced primary side to a secondary side referenced to the source voltage of the high-side MOSFET to power the MOSFET driver. In addition to a floating power supply such as the bootstrap power supply described above or a transformer-based power supply, level-shifting circuitry is typically required to convert the control signals input to the MOSFET driver to the desired values, such as a voltage range referenced to the source, and in some cases, to provide isolation. Such level shifting is typically achieved using a capacitively coupled transformer or a level-shifting transformer.

[0005] The floating power supply and level shifting of a power switch driver may require complex and physically large circuitry, which is not easily integrated into a semiconductor die. A simplified circuit topology for driving the power switch is desired, where most or all of the circuitry can be integrated into the semiconductor die. Summary of the Invention

[0006] This invention discloses circuit topologies and related techniques for driving power switches. These topologies utilize the technique of transmitting power and switch control signals using the same transformer. Such circuits are suitable for integration within a semiconductor die and, in many cases, can avoid large capacitors and magnetic devices (e.g., transformer cores). The resulting power switch drivers are typically smaller, simpler, and cheaper than existing drivers, and much of the driver circuitry can be integrated with the power switch within the same semiconductor die.

[0007] According to an embodiment of a transformer-based switching device, the switching device includes a power transistor, a transformer, and a rectifier. The power transistor has a first load terminal and a second load terminal, and a control terminal for controlling conduction between the first and second load terminals. The transformer has a primary winding and a secondary winding. A rectifier is coupled between the secondary winding and the control terminal. The secondary winding and the rectifier are configured to transmit the required energy and control timing for driving the control terminal based on an input waveform coupled to the primary winding, in order to control conduction between the first and second load terminals. During the power transistor's on-interval, the input waveform has multiple high-frequency pulses, which the transformer and rectifier convert into on-control pulses provided to the control terminal to turn on the power transistor. During the power transistor's off-interval, the input waveform has a direct current (DC) level.

[0008] According to an embodiment of a transformer-driven power switching system, the system includes a power transistor, a transformer, a rectifier, and control logic circuitry. The power transistor has a first load terminal and a second load terminal, and a control terminal for controlling conduction between the first and second load terminals. The transformer has a primary winding and a secondary winding. The rectifier couples the secondary winding of the transformer to the control terminal. The control logic circuitry includes a control input and a frequency generator. The frequency generator is configured to provide a high-frequency on-state signal to the primary winding in response to receiving an on-state level at the control input, and to provide a direct current (DC) signal to the primary winding in response to receiving an off-state level at the control input.

[0009] Other features and advantages will be recognized by those skilled in the art upon reading the following detailed description and viewing the accompanying drawings. Attached Figure Description

[0010] The elements in the accompanying drawings are not necessarily proportional to each other. Similar reference numerals denote corresponding similar parts. Features of the various illustrated embodiments can be combined unless they are mutually exclusive. Embodiments are depicted in the accompanying drawings and described in detail below.

[0011] Figure 1 A high-level schematic diagram of a transformer-driven power switching system is shown.

[0012] Figure 2A A circuit topology diagram of a transformer-driven power switching device using a normally closed pull-down switch and a pull-down resistor is shown. Figure 2B A circuit for providing a diode based on a GaN switch is shown.

[0013] Figure 3 It shows the corresponding Figure 2A The voltage and current waveforms of the circuit.

[0014] Figure 4 A schematic diagram of another circuit topology for a transformer-driven power switching device using a normally closed pull-down switch is shown, in which... Figure 2A The pull-down resistors were replaced by a diode bridge.

[0015] Figure 5 It shows the corresponding Figure 4 The voltage and current waveforms of the circuit.

[0016] Figure 6 A schematic diagram of another circuit topology for a transformer-driven power switching device using a normally closed pull-down switch is shown.

[0017] Figure 7 It shows the corresponding Figure 6 The voltage and current waveforms of the circuit.

[0018] Figure 8 A schematic diagram of another circuit topology for a transformer-driven power switching device using a normally closed pull-down switch is shown.

[0019] Figure 9 It shows the corresponding Figure 8 The voltage and current waveforms of the circuit.

[0020] Figure 10 A schematic diagram of the circuit topology of a transformer-driven power switching device using a normally open pull-down switch and a pull-down resistor is shown.

[0021] Figure 11 It shows the corresponding Figure 10 The voltage and current waveforms of the circuit.

[0022] Figure 12 A schematic diagram of another circuit topology for a transformer-driven power switching device using a normally open pull-down switch and a pull-down resistor is shown.

[0023] Figure 13A A schematic diagram of the circuit topology of a transformer-driven power switching device using a normally open pull-down switch and a voltage clamp is shown. Figure 13B An example of a GaN-based voltage clamper is shown.

[0024] Figure 14 It shows the corresponding Figure 13A The voltage and current waveforms of the circuit.

[0025] Figure 15 A schematic diagram of the circuit topology of a transformer-driven power switching device using a normally open pull-down switch with passive discharge is shown.

[0026] Figure 16 It shows the corresponding Figure 15 The voltage and current waveforms of the circuit.

[0027] Figure 17 A schematic diagram of the circuit topology of a transformer-driven power switching device without a pull-down switch is shown.

[0028] Figure 18 It shows the corresponding Figure 17 The voltage and current waveforms of the circuit.

[0029] Figure 19 A schematic diagram of the circuit topology of a transformer-driven power switching device that does not include a pull-down switch and uses a rectifier bridge to directly drive the gate of the power switch is shown.

[0030] Figure 20 It shows the corresponding Figure 19 The voltage and current waveforms of the circuit.

[0031] Figure 21 A schematic diagram of the circuit topology of a transformer-driven power switching device that does not include a pull-down switch and passively discharges the gate of the power switch is shown.

[0032] Figure 22 It shows the corresponding Figure 21 The voltage and current waveforms of the circuit.

[0033] Figure 23 A system-level package including a transformer-driven power switching device is shown.

[0034] Figure 24 A schematic diagram of the circuit topology of a transformer-driven bidirectional power switch is shown.

[0035] Figure 25 A schematic diagram of the circuit topology of a transformer-driven power switching device, including a normally open power switch, is shown.

[0036] Figure 26 It shows the corresponding Figure 25 The voltage and current waveforms of the circuit.

[0037] Figure 27 A schematic diagram of the circuit topology of a transformer-driven bidirectional switching device, including a normally open power switch, is shown.

[0038] Figure 28 A schematic diagram of the circuit topology of a transformer-driven switching device including power MOSFETs is shown.

[0039] Figure 29 It shows the corresponding Figure 28 The voltage and current waveforms of the circuit. Detailed Implementation

[0040] The embodiments described herein disclose circuitry and techniques for transformer-driven power switching devices and associated systems. These embodiments are primarily described in the context of power switches within transformer-driven power switching devices that are based on gallium nitride (GaN) gate-injected transistors (GITs). The circuit topologies described herein offer particular advantages when GaN transistors are used for power switching. Many of these advantages relate to the fact that GaN transistors require relatively low energy transfer (charge) to or from their gate during transitions between on and off states, at least compared to conventional metal-oxide-semiconductor field-effect transistors (MOSFETs) or similar transistors. This relatively low charge requirement makes it possible for coreless transformers to provide the energy (charge) needed to turn the power switch on and / or off, while avoiding the bulky size and cost of the magnets used in cored transformers typically required by other switching types, such as MOSFETs. Coreless transformers can be integrated within semiconductor dies, printed circuit boards (PCBs), etc., to achieve small size and low cost.

[0041] In addition to the transformer described above that transfers alternating current (AC) energy from the primary side to the secondary side, transformer-driven switching devices also include a rectifier for converting the secondary-side AC energy into direct current (DC) energy suitable for application to the gate of the power switch. When the power switch is based on GaN transistors such as GITs, GaN dies can also advantageously be used to provide other circuit elements. Diodes, resistors, and even capacitors that are part of the rectifier or other driver circuitry can be integrated with the power switch into the same GaN die, providing a high level of integration and associated small size and low cost.

[0042] Although GaN-based GITs, which are enhancement-mode (normally off) high electron mobility transistors (HEMTs), are used in many examples in this paper, other HEMTs, including depletion-mode (normally on) HEMTs and HEMTs based on other Group III / V materials, offer similar advantages to those described above, which will be further explained below. However, it should be understood that the circuit topologies and techniques described herein are not limited to use with GaN-based power switches or other HEMT power switches, and additional advantages can be achieved when these topologies and techniques are used with other types of power switches, including, for example, silicon-based power switches and silicon carbide-based power switches.

[0043] The invention is described below using several exemplary circuits, which should not be considered limiting. Figure 1A block diagram of a system-in-package (SiP) based on an enhancement-mode GaN transistor (GIT) is shown. Several detailed circuits are described below, in which rectifiers and other driver circuitry are integrated with the GIT within the GaN die. These circuits vary primarily based on how the gate of the GIT is pulled down to turn it off. For simplicity, these circuits are categorized as those using normally-off pull-down driver switches, those using normally-on pull-down driver switches, and those not using active pull-down driver switches. Following these circuit descriptions, example circuits for driving a bidirectional GaN power switch including a GIT, a normally-on GaN power switch, and a bidirectional switch including a normally-on GaN power switch are shown. Next, an example of a transformer-based switching device utilizing a power MOSFET instead of a GaN-based power switch is described.

[0044] Figure 1 A system-in-package (SIP) 100 of a transformer-driven power switching device is shown. System 100 includes a GaN power switch Q1 and circuitry for driving the GaN power switch Q1. System 100 has load terminals 101, 102 corresponding to the drain (D) and source (S) of power switch Q1, and an input terminal 104 for receiving control signals from an external controller (not shown for illustration). As shown, the control signal is a pulse-width modulation (PWM) waveform, where, for example, the duty cycle or frequency of the PWM waveform controls when the power switch is turned on. For example, power switch Q1 may be turned on during high intervals of the PWM waveform and turned off during low intervals of the PWM waveform. In the illustrated embodiment, the PWM waveform is a ground-referenced digital signal alternating between 0V and CMOS / TTL voltage levels. System 100 also has a terminal 106 for connecting to a power supply providing a voltage rail Vcc and a terminal 107 for grounding.

[0045] System 100 includes a transformer and control logic 110, as well as a GaN die 140. The transformer and control logic 110 includes a drive controller 120 and a transformer 130. The drive controller 120 includes ground-based control logic. A signal generator 122 within the drive controller 120 provides a high-frequency signal to the transformer 130. The frequency of the provided signal is significantly higher than the frequency of the PWM waveform input to system 100. For example, the PWM signal may switch the power switch at 150 kHz, while the signal generator 122 provides a square wave with a frequency greater than 100 MHz, such as a 200 MHz square wave. In the exemplary embodiment described in further detail below, when the PWM input is high, the high-frequency signal from the signal generator 122 is coupled to the transformer 130, and when the PWM input is low, no power is supplied to the transformer 130. The interval at which no power is supplied to the transformer 130 can be set by coupling 0V or some other DC level to the transformer when the PWM input is low, or by disconnecting the transformer 130 to float its input. In another implementation described below as an example, an additional transformer winding may be used. Figure 1 (Not shown in the image) to actively turn off the power switch Q1.

[0046] The signal generator 122 can be implemented using a half-bridge circuit, which switches at a high frequency to provide voltage to the transformer 130 at the voltage rail V. CC A high-frequency waveform alternating between +Vcc and -Vcc. In some implementations, as shown in the following examples, signal generator 122 includes a full-bridge circuit to provide a signal alternating between +Vcc and -Vcc. Other signal generation techniques may also be used. Because signal generators are well known in the art, further details are not provided.

[0047] Transformer 130 includes a primary winding 132 and a secondary winding 134. Some embodiments also include a magnetic core 136, shown in dashed lines to indicate that the magnetic core 136 is optional. In a preferred embodiment including typical circuitry, where the power switch Q1 is a GaN-based transistor, transformer 130 is coreless, i.e., without a magnetic core 136. (As used herein, the term "coreless transformer" describes a transformer in which the windings are not coupled by a magnetic core. Such coreless transformers may include air cores, etc.) For embodiments with such coreless transformers, drive controller 120 and transformer 130 may be integrated within the same semiconductor die, such as a silicon die. In other embodiments, a cored transformer 130 may be required, and the transformer and control logic 110 may or may not be implemented within a single semiconductor die. For example, a conventional MOSFET power switch may require more gate charge over a period of time than a coreless transformer can practicably provide, thus potentially necessitating a cored transformer instead. In such a case, it may be impossible to implement the transformer and control logic 110 within a single semiconductor die.

[0048] GaN die 140 includes a GIT power switch Q1, a rectifier 150, and a pull-down circuit 160. Rectifier 150 converts AC energy supplied by secondary winding 134 into a DC voltage suitable for turning on the gate (G) of power switch Q1. For example, rectifier 150 can convert AC energy from transformer 130 into a DC gate-source voltage V greater than the threshold voltage (e.g., 3.3V) of power switch Q1. GS This is to enable the power switch Q1 to be turned on. Pull-down circuit 160 is used to control the gate voltage V. GS Discharge so that power switch Q1 is turned off when the transformer and control logic 110 no longer provide AC energy. Such discharge can be achieved using a pull-down switch or other active components, or it can be achieved using passive components.

[0049] Although system 100 includes a primary winding, a secondary winding, and a rectifier, several circuit topologies described below include additional windings and additional rectifiers. Variations of transformer-driven power switching devices are provided below, wherein the described circuits differ primarily in the number of transformer windings, rectifier circuitry, and pull-down circuitry.

[0050] Transformer-based switching device based on normally closed pull-down switch

[0051] Figure 2AThe circuit diagram of a transformer-driven switching device 200 is shown. This device includes a set of transformer windings and a first rectifier for turning on a power switch Q1. It also includes another set of transformer windings, a second rectifier, and a normally closed pull-down switch for turning off the power switch Q1. As shown, the transformer-driven switching device is divided into a transformer 230 and a GaN die 240. (The above information is incomplete and requires further context.) Figure 1 As described, transformer 230 may be part of a silicon die that also includes a drive controller.

[0052] Transformer 230 includes a first primary winding 232a, a first secondary winding 234a, and (optionally) a first core 236a for providing secondary-side AC current I. SEC_ON The transformer 230 also includes a second primary winding 232b, a second secondary winding 234b, and (optionally) a second core 236b for providing secondary-side AC current I. SEC_OFF Signal generators such as Figure 1 Signal generator 122 generates a primary-side voltage V coupled to the primary windings 232a and 232b. PRIM_ON and V PRIM_OFF For example, the signal generator can provide a substantially continuous AC waveform, wherein the AC waveform is coupled to the first primary winding 232a up to the on interval of the power switch Q1, and alternatively coupled to the second primary winding 232b up to the off interval of the power switch Q1. In another example, separate signal generators can be used to drive the primary windings 232a, 232b, wherein each winding is activated / deactivated as needed.

[0053] In addition to the power switch Q1, the GaN die 240 also includes a turn-on rectifier (diode bridge) 250, a turn-off rectifier 252, and a pull-down circuit 260. The turn-on rectifier 250 controls the secondary side current I. SEC_ON Rectification is performed to provide a positive DC current I to the gate (G) of the power switch. GS The rectifier 250 maintains a positive voltage V at its gate (G) during the turn-on interval. GS This prevents the gate voltage from discharging back to transformer 230. A capacitor may be included to allow the gate voltage V to... GS Smooth, however, as shown in the figure, the gate capacitance of power switch Q1 provides filtering.

[0054] When the off signal V PRIM_OFF When applied to the second primary winding 232b, the pull-down circuit 260 and the turn-off rectifier 252 are used to turn off the power switch Q1. The turn-off rectifier 252 controls the resulting secondary-side current I. SEC_OFFThe current is rectified to provide charging current to the gate of pull-down driver switch Q2, thereby turning on pull-down driver switch Q2. With switch Q2 on, the gate of power switch Q1 discharges, and its gate-source voltage V... GS The drive value is zero, thus turning off the power switch Q1.

[0055] It is important to ensure that the normally closed pull-down driver switch Q2 remains completely off throughout the entire ON state of power switch Q1. When the secondary winding 234b does not supply power, for example, when power switch Q1 is ON, the pull-down resistor R... PD (Shown as a 1kΩ resistor) Keeps the pull-down driver switch Q2 off.

[0056] As previously suggested, pull-down driver switch Q2, resistor R PD Rectifiers 250 and 252 are preferably implemented within the same GaN die as power switch Q1. Normally closed pull-down driver switch Q2 is constructed in almost the same manner as power switch Q1, but can be smaller because it does not need to support the high current level required by power switch Q1. For example, pull-down driver switch Q2 may include much less GaN channel (finger) than power switch Q1, may have a smaller channel width compared to power switch Q1, and / or may omit or have a shorter drift (voltage blocking) region than power switch Q1. Pull-down resistor R can be implemented within GaN die 240 using one or more two-dimensional electron gas (2DEG) regions of the GaN semiconductor die. PD Each of the regions is essentially a gateless GaN HEMT.

[0057] The diodes within rectifiers 250 and 252 can be implemented within a GaN die using GaN transistors configured as gate-controlled diodes or PN diodes. Although in Figure 2A Not shown in the diagram, but in some embodiments, these diodes may be GaN Schottky diodes. For example... Figure 2B As shown, a gate-controlled diode 280 can be constructed by coupling the gate of a normally open GaN-based switch (GIT) to its source, thereby converting the switch into a two-terminal device (diode) in which the gate / source acts as the anode and the drain as the cathode. Such a gate-controlled diode typically has a threshold (knock-off) voltage of 0.9V to 1.5V. Alternatively, a PN diode 282 can be constructed by coupling the drain and source of the GIT together, thereby converting the switch into a two-terminal device (diode) in which the gate is the anode and the drain / source is the cathode. Such a PN diode has a relatively stable threshold (knock-off) voltage of approximately 3.3V. Unless otherwise specified, the GaN-based diodes (including...) in the examples herein... Figure 1The rectifiers 250 and 252 are gate-controlled diodes with a threshold voltage of 1.2V.

[0058] like Figure 2A As shown, each winding in transformer 230 has a turns ratio of 1:2, but other ratios can be used. For a primary-side power supply Vcc of 3V, where the primary-side AC voltage V... PRIM_ON V PRIM_OFF Alternating between +3V and -3V, a 1:2 turns ratio ensures sufficient turn-on voltage, providing, for example, a voltage greater than 3.3V at the respective gates of power switch Q1 and pull-down driver switch Q2. The 3V primary-side voltage is doubled by transformer 230 and reduced by the diode voltage drop (e.g., 2 * 1.2V = 2.4V) within rectifiers 250 and 252. Furthermore, both the primary-side and secondary-side voltages are slightly reduced by the resistive losses of transformer 230.

[0059] Figure 3 It shows the corresponding Figure 2A The voltage and current waveforms of the transformer-driven power switching device. The primary side AC voltage V is driven at a frequency of 200MHz. PRIM_ON V PRIM_OFF And it switches power switch Q1 at a frequency of 150kHz. For example... Figure 3 As shown, between time t0 and t1, a 200MHz signal with a + / -3V voltage is used to actively drive the turn-on voltage V. PRIM_ON And the corresponding AC current I PRIM_ON It is supplied to the first primary winding 232a. The resulting secondary side AC current I SEC_ON The current is output from the secondary winding 234a and rectified by the diode bridge 250. The diode bridge 250 outputs an average rectified current I of approximately 16 mARMS. GS During the on-off interval, it charges the gate (G) of power switch Q1. The resulting gate-source voltage V is close to 3.6V. GS The voltage exceeds the turn-on threshold of power switch Q1 (e.g., 3.3V), causing power switch Q1 to turn on during these intervals. (Transformer 230 boosts the 3V input to approximately 6V, then reduces it by 2.4V due to the voltage drop across the two gate-controlled diodes within rectifier 250.)

[0060] During the interval from time t1 to t2, the voltage V is turned on. PRIM_ON Ineffective, such as floating or holding at 0V, resulting in no primary-side current I. PRIM_ON or secondary side current I SEC_ON Flow. During this interval, the turn-off voltage V is actively driven by alternating 200MHz signals between +3V and -3V. PRIM_OFFAnd the corresponding turn-off current I PRIM_OFF This is supplied to the second primary winding 232b. The resulting secondary-side AC current I... SEC_OFF The output from the secondary winding 234b, rectified by the diode bridge 252, drives the gate of the pull-down switch Q2 to approximately 3.6V, such as V. PD_GS As shown in the waveform, it is higher than the turn-on threshold of pull-down switch Q2. When pull-down driver switch Q2 is turned on, the gate and source of power switch Q1 are short-circuited together, thus turning off power switch Q1.

[0061] At time t2, the voltage V is actively driven to turn on again. PRIM_ON This is to allow the power switch Q1 to be switched on again. Pull-down resistor R PD This is used to passively pull the gate of the pull-down driver switch Q2 down to the supply voltage, thereby forcing switch Q2 to turn off. This can be achieved starting from time t2 at voltage V. PD_GS This can be seen from the attenuation.

[0062] Figure 4 It shows Figure 2A Variation 400 of the transformer-driven switching device, wherein the passive pull-down resistor R PD Instead, a stacked diode bridge 460 is used to actively drive the pull-down driver switch Q2 to turn off when the power switch Q1 is turned on. The diode bridge 460 is driven by another secondary winding 435a, which is also coupled to (and driven by) the first primary-side winding 432a. This diode bridge 460 is stacked in the opposite direction to the rectifier 252 and has twice the number of diodes in each branch to drive a negative voltage to the gate of the pull-down driver switch Q2 when the power switch Q1 is turned on. Since both secondary windings 434a and 435a are coupled to the same (on) primary winding 432a, energy is transferred simultaneously, ensuring that the pull-down driver switch Q2 is driven off when the power switch Q1 is turned on. The stacked diode bridge is configured to discharge the gate of the pull-down driver switch Q2 and not only the gate voltage V. PD_GS Driven below the threshold voltage of switch Q2, and the gate voltage V PD_GS It can also be driven by a negative voltage. Driving a negative voltage to the gate of switch Q2 can accelerate the turn-off of switch Q2, thereby accelerating the turn-on transition of power switch Q1.

[0063] As in Figure 2A Similar to the switching device, the pull-down diode bridge (rectifier) ​​252 is used to drive a positive voltage onto the pull-down gate so that the pull-down switch Q2 is turned on when the power switch Q1 is turned off. The activation voltage of the diode bridge 252 must be lower than the activation voltage of the diode bridge 460. Figure 4Transformer 430 and Figure 2A The transformer differs in that the turns ratio of the first primary winding 432a to the secondary windings 434a and 435a is 1:1.2:1:2, and the turns ratio of the second primary winding 432b to its secondary winding 434b is 1:1. This configuration ensures that the voltage driven to the diode bridge 460 is sufficient to activate the diodes when the power switch Q1 is turned on.

[0064] Figure 5 It shows the corresponding Figure 4 The voltage and current waveforms of a transformer-driven power switching device. For example... Figure 3 As shown, the primary-side AC voltage V is driven at a frequency of 200MHz. PRIM_ON V PRIM_OFF The power switch Q1 is switched at a frequency of 150kHz. The primary side turn-on voltage and the primary side turn-off voltage are related to... Figure 3 The voltage levels alternate between +5V and -5V. Besides their voltage levels, the primary side current I... PRIM_ON I PRIM_OFF and secondary side current I SEC_ON I SEC_OFF and Figure 3 Similar. With Figure 3 different, Figure 5 It shows the relationship with pull-down current I SEC_PD The corresponding waveform indicates that the pull-down current originates from... Figure 4 The output current of the secondary winding 435A. This current is exactly at the on-voltage V. PRIM_ON After becoming active, the current rises, causing the gate of pull-down switch Q2 to discharge. Pull-down current I SEC_PD It decays toward zero as the gate charge dissipates. Furthermore, Figure 5 This shows the pull-down gate voltage V when power switch Q1 is turned on. PD_GS It can decrease to a negative value.

[0065] although Figure 4 The switching device relative to Figure 2A Switching devices can offer some performance advantages, such as faster connection speeds, but... Figure 4 The switching device is more complex because it requires another secondary winding and another stack of diode bridges. Figure 6 The switching device has similar advantages, but with lower circuit complexity.

[0066] Figure 6 Another version 600 of the transformer-driven switching device is shown. This version 600 is similar to... Figure 2A The switching device, except that the pull-down circuit 660 includes a pull-down diode D for driving the pull-down switch Q2 to its off state. PD Instead of including Figure 2A The pull-down resistor R of the 260 pull-down circuit PD In addition, this topology offers the advantage of using the same current loop 690 that turns on power switch Q1 to drive pull-down switch Q2 off. This is achieved by using pull-down diode D... PD Connected between the source and gate of pull-down switch Q2, diode D... PD The threshold voltage will provide a negative voltage at the gate of the pull-down switch Q2, thereby ensuring that the pull-down switch Q2 is driven to its off state whenever the main switch is driven to turn on.

[0067] Figure 7 It shows the corresponding Figure 6 The voltage and current waveforms of the switching device 600. These waveforms are similar to... Figure 3 The waveform. However, during the interval when power switch Q1 is turned on, the gate-source pull-down voltage V... PD_GS It is driven to a negative value (approximately -1.2V, corresponding to the forward bias voltage of the gate-controlled diode). Additionally, the rectified current I driven to the gate of power switch Q1... GS (approximately 8 mARMS) lower than Figure 3 The corresponding current in the circuit is because there is no pull-down resistor in the pull-down circuit 660 that discharges a portion of the steady-state current.

[0068] Figure 8 Another transformer-driven switching device 800 is shown, wherein the pull-down circuit 860 includes a cascaded pull-down switch Q2 and does not require a pull-down resistor or diode. The gate of the pull-down switch Q2 is connected to the source of the power switch Q1. A turn-off diode bridge (rectifier) ​​252 is connected across the gate and source of the pull-down switch Q2. As in the switching device described above, the drain of the pull-down switch Q2 is connected to the gate of the power switch Q1, but the source of the pull-down switch Q2 is not connected to the source of the power switch Q1. With this configuration, the power switch Q1 is turned on in a manner similar to that described for the previous switching device, but the turn-off differs in that a negative voltage is applied to the gate of the power switch Q1 during the turn-off interval. When the turn-off diode bridge 252 is activated, a positive pull-down gate-source voltage V is generated. PD_GS This is to turn on the pull-down switch Q2. With the drain and source of the pull-down switch Q2 short-circuited together, the gate of the power switch Q1 is connected to the negative side of the turn-off rectifier 252, while the source of the power switch Q1 is connected to the positive side of the turn-off diode bridge 252. Therefore, the voltage generated by the turn-off rectifier 252 provides a negative voltage across the gate-source terminals of the power switch Q1, where the magnitude of this negative voltage is the same as the magnitude of the positive voltage applied to the gate and source terminals of the pull-down switch Q2. That is, when the power switch Q1 is off, V...GS =-V PD_GS In addition to providing a relatively fast turn-off transition, the negative voltage at the gate of power switch Q1 has another benefit: it keeps power switch Q1 firmly in its off state, thereby reducing the risk of accidental turn-on of power switch Q1 due to noise or ringing.

[0069] Figure 9 It shows the corresponding Figure 8 The voltage and current waveforms of the transformer-driven switching device 800 are shown. These waveforms are similar to those corresponding to the previously described switching device, but the notable exception is the gate-source voltage V of power switch Q1 when it is driven to its off state. GS It is driven to a negative value of approximately -3.3V. This voltage is driven to the gate-source voltage V of the pull-down switch Q2 during the off state of power switch Q1. PD_GS The negative value (e.g., +3.3V).

[0070] Transformer-based switching device based on normally open pull-down switch

[0071] The transformer-based switching device in the previous section used a normally-off switch to actively pull down the power switch control terminal (gate). The transformer-based switching device described below alternatively uses a normally-open pull-down switch. Although other types of normally-open switches can be used, the following examples use a normally-open GaN switch. Except for the absence of the p-type gate structure typically included within the GIT, such a switch is constructed in almost the same way as a normally-off GIT. Such a normally-open GaN switch can be a HEMT in its pristine state, such as a depletion-type HEMT. Because normally-open HEMTs can be constructed using the same material as power switches, such as GaN, they can be integrated with the power switch in the GaN die in the same way that normally-off pull-down switches might be. The normally-open GaN switch used in the following examples is turned on (operated) when no voltage is applied or when a voltage above the switching threshold is applied to the switch gate. The turn-off threshold is typically a negative voltage in the range of -4V to -7V. When a gate-to-source voltage below the turn-off threshold is applied, the normally-open pull-down switch is turned off.

[0072] Figure 10 The circuit diagram of a transformer-driven switching device 1000 using a normally open pull-down switch Q2 is shown. The transformer-driven switching device 1000 includes devices having each of similar characteristics. Figure 2A The corresponding components shown are the power switch Q1, the GaN die 1040 for the rectifier 250 and the pull-down rectifier 252. The transformer-driven switching device 1000 also includes a pull-down circuit 1060, which includes a normally open pull-down switch Q2 and a pull-down resistor R. PDCompared to the aforementioned circuit, the pull-down rectifier 252 is configured to apply a negative voltage to the gate of the pull-down switch Q2 relative to the source of the pull-down switch Q2 during the interval when the power switch Q1 is on. During the off interval of the power switch Q1, the resistor R... PD The charge between the gate and source of pull-down switch Q2 is balanced, thereby pulling down the gate-source voltage V. PD_GS The drive is zero, thus connecting the normally open pull-down switch Q2.

[0073] By using a normally open pull-down switch Q2 instead of a normally closed pull-down switch, power switch Q1 is in the off (non-conducting) state by default. Pull-down switch Q2 short-circuits the gate and source of power switch Q1 unless its gate is driven with a sufficiently negative voltage. This default state provides another benefit: power switch Q1 is not turned on during startup or in the absence of power. This, in turn, prevents shoot-through problems when power switch Q1 is part of a half-bridge or similar configuration, thus improving system safety and reliability. This fail-safe operation is achieved without the complex circuitry required to apply a negative gate voltage to the power switch before system startup, as is typical in GaN driver circuitry. Other normally open pull-down switch-based circuits described below also offer these benefits.

[0074] Figure 10 The normally open pull-down switch Q2 has a turn-off threshold of -4V or lower (a larger negative value), that is, when a gate-source voltage V less than -4V or lower is applied... PD_GS When the voltage is low, switch Q2 is off; otherwise, it is on. To obtain such a negative voltage, transformer 1030 includes a primary winding 1032b and a secondary winding 1034b with a turns ratio of approximately 1:3 or higher, which differs from previously described transformers, such as... Figure 2A The turns ratio of transformer 230 and windings 232a, 234a (e.g., 1:2). The turn-off threshold of pull-down switch Q2 may vary due to manufacturing differences or due to different implementations / types of pull-down switch Q2. The transformer turns ratio and / or the type and number of diodes in pull-down rectifier 252 can be adapted to accommodate different turn-off thresholds of pull-down switch Q2.

[0075] Figure 11 It shows the corresponding Figure 10 The voltage and current waveforms of the transformer-driven switching device 1000 are shown. The primary-side voltage V is the same. PRIM_ON It is applied to the two primary-side windings 232a and 1032b. This voltage V PRIM_ON It is effective during the on-state interval of power switch Q1 (alternating between approximately +3V and -3V), and ineffective during the off-state interval of power switch Q1 (approximately 0V). Figure 11As shown, the secondary side current waveform I SEC_ON I SEC_PD_OFF Their timing sequences are similar, but due to different transformer winding turns ratios (e.g., 1:2 vs. 1:4), their current levels differ slightly during the switching interval of power switch Q1. When the primary side voltage V PRIM_ON When it is effective and power switch Q1 is turned on, such as the voltage waveform V PD_GS As shown, the gate-source voltage V of the pull-down switch Q2 PD_GS It is driven by a negative voltage, such as -9V. This negative voltage keeps the pull-down switch Q2 in its off state. When the primary side voltage V... PRIM_ON When the resistor R is inactive (e.g., approximately 0V) and the power switch Q1 is off, the resistor R... PD Voltage V PD_GS Pulling the voltage down to 0V turns on the pull-down switch Q2 and short-circuits the gate and source terminals of the power switch Q1, thereby keeping the power switch Q1 in the off state.

[0076] Primary windings 232a and 1032b receive the same input, and the "off" signal is essentially just the absence of an "on" signal, such as... Figure 11 waveform V PRIM_ON As shown in the diagram. This means that no power is required to maintain the off state of the power switch Q1, unlike the circuit previously described using a normally closed pull-down switch. Relatedly, the power consumption of the transformer-driven switching device 1000 depends on the duty cycle of the power switch Q1. Furthermore, because the same signal is applied to both primary windings, the transformer 1030 can be simplified to a single primary winding coupled to both secondary windings. Figure 12 A transformer-driven switching device 1200 is shown. Figure 12 The transformer 1230 includes a single primary winding 1232a coupled to two secondary windings 1234a and 1234b. A magnetic core 1236a may be included to assist coupling, or the core 1236a may be omitted. The turns ratios from the primary winding 1232a to the secondary windings 1234a and 1234b are 1:2 and 1:4, respectively. In other aspects, Figure 12 Circuit 1200 and Figure 10 The circuit is the same as 1000, and the voltage and current waveforms of circuit 1200 are similar. Figure 11 The voltage and current waveforms are shown.

[0077] Figure 13A A transformer-driven switching device 1300 is shown, which uses a normally open pull-down switch Q2, but the passive discharge of the gate of switch Q2 is replaced by active discharge. This device 1300 is similar to... Figure 4Device 400 also enables the gate of the pull-down switch to actively discharge and provides similar benefits to device 400. Device 1300 includes a transformer 1330 and a GaN die 1340.

[0078] Transformer 1330 is similar to others except that it can have different turns ratios. Figure 2A The transformer 230, for example, has a turns ratio of 1:5 between the first windings 1332a and 1334a, and a turns ratio of 1:3 between the windings 1332b and 1334b. The GaN die 1340 includes each of these components similar to... Figure 10 The corresponding components in the device 1000 are the turn-on rectifier 250, the turn-off rectifier 252, and the power switch Q1. The GaN die 1340 further includes a pull-down circuit 1360 with a normally open pull-down switch Q2 and a voltage clamp 1362. The voltage clamp 1362 functions similarly to a diode with a large forward bias threshold (e.g., 9V), thereby generating a voltage drop V from the source to the gate of the pull-down switch Q2. CL When voltage clamp 1362 is activated, i.e., when sufficient voltage is provided to forward bias voltage clamp 1362, a negative voltage is driven relative to the pull-down source to its gate to hold pull-down switch Q2 in its off state. For example, if voltage clamp 1362 is provided with a +9V bias voltage, then when voltage clamp 1362 is forward biased, V PD_GS = -9V. Since the turn-off threshold of a normally open switch is typically -4V to -7V, the pull-down gate-source voltage V PD_GS =9V will robustly keep the pull-down switch Q2 off.

[0079] Voltage clamp 1362 is part of the on-current loop 1390, such that when power switch Q1 is turned on, voltage clamp 1362 forces pull-down switch Q2 to its off state. A sufficiently high voltage from the on-rectifier 250 will turn on power switch Q1 and forward bias voltage clamp 1362, thereby turning off pull-down switch Q2. For a primary-side input V alternating between +3V and -3V... PRIM_ON Furthermore, the transformer turns ratio is 1:5, and the amplitude of the secondary voltage supplied to the rectifier 250 is 15V. Assuming the voltage drop within the rectifier 250 is 2.4V, the resulting rectifier output voltage is 12.6V, which is approximated by the clamping voltage V. CL (e.g., 9V) and the gate-source voltage V of power switch Q1 GS (e.g., between 3.6V).

[0080] When power switch Q1 is off, the shutdown rectifier 252 applies a voltage to pull-down switch Q2, thereby turning on pull-down switch Q2 and short-circuiting the gate and source of power switch Q1. (Any voltage above the threshold voltage, such as -4V, turns on pull-down switch Q2.) A second voltage clamp 1364 is included in the shutdown current loop to prevent activation of the shutdown rectifier 252 during the on-interval of power switch Q1. (Without the second voltage clamp 1364, current would flow through the shutdown rectifier 252 and bypass the voltage clamp 1364 during the on-interval.) For a primary-side input of + / -3V and a turns ratio of 1:4, the resulting secondary-side voltage supplied to the shutdown rectifier 252 is + / -12V. Assuming that turning off rectifier 252 produces a voltage drop of 2.4V, the resulting +9.6V voltage forward biases the second voltage clamp 1364, which has a clamping voltage V. CL (e.g., 9V), and provides, for example, a pull-down gate-source voltage V. PD_GS =0.6V. This voltage turns on the pull-down switch Q2, which in turn turns off the power switch Q1.

[0081] As previously mentioned Figure 2B As described, gate-controlled diodes and PN diodes can be constructed from GaN transistors. Voltage clamps 1362 and 1364 may include a series stack of such diodes. Figure 13B Examples of voltage clamps 1362a, 1362b, and 1362c based on gate-controlled diodes and / or PN diodes are shown. The first voltage clamp 1362a includes a plurality of gate-controlled diodes connected in series. Each such gate-controlled diode typically has a knee (threshold) voltage of 0.9V to 1.5V. For example, if N=6 in the first voltage clamp 1362a and the threshold voltage is 1.5V, then the first voltage clamp 1362a will provide a clamping voltage V. CL =9V. The illustrated second voltage clamp 1362b includes three PN diodes connected in series, each of which has a 3V inflection point (threshold) voltage. Therefore, the illustrated second voltage clamp 1362b will provide a clamping voltage V. CL =9V. Figure 13A Voltage clamps 1362 and 1364 each have a clamping voltage of 9V, therefore the circuitry of the second voltage clamp 1362b can be used. The third voltage clamp 1362c shown comprises a combination of two PN diodes and a gate-controlled diode. Using the aforementioned representative threshold voltage, the third voltage clamp 1362c provides a clamping voltage V. CL=7.2V. Other combinations of gate-controlled diodes and PN diodes can be assembled to achieve the desired clamping voltage level, for example, to provide a clamping voltage V corresponding to the turn-off threshold of a specific normally open pull-down switch. CL .

[0082] Figure 14 It shows the corresponding Figure 13A The voltage and current waveforms of the transformer-driven power switching device 1300 are shown. As in the previous example, during the on-time interval of power switch Q1, the primary side on-signal V... PRIM_ON The voltage alternates between +3V and -3V. After passing through transformer 1330 (turns ratio 1:5) and rectifier 250, the resulting 15V provides a gate-source voltage V of 3.6V. GS and -9V pull-down gate-source voltage V PD_GS As shown in the figure. During the off-time interval of power switch Q1, the primary side off signal V... PRIM_OFF The voltage alternates between +3V and -3V. After passing through transformer 1330 (turns ratio 1:4) and turning off rectifier 252, the resulting 9.6V forward biases the second voltage clamp 1364 and provides a near-zero pull-down gate-source voltage V. PD_GS As shown in the figure. This turns on the pull-down switch Q2, as shown in the figure, and reduces the gate-source voltage V of the power switch Q1. GS Drive to 0V.

[0083] While the switching device 1300 has the advantage of providing active on / off control of the power switch Q1 and associated control of the on / off switching speed, the switching device 1300 includes two sets of transformer windings, two voltage clamps, and two rectifiers. For applications that do not require active control of the power switch's off-time, a simpler circuit providing fail-safe protection may be desirable. Figure 15 A simpler circuit like this is shown.

[0084] Figure 15 It shows the relationship with Figure 13A A similar transformer-driven switching device 1500, except that the active discharge provided by the turn-off rectifier 252 is replaced by a passive discharge provided by a pull-down resistor, thus simplifying the circuit. In addition to eliminating... Figure 13A In addition to shutting off the rectifier 252, this passive discharge enables the removal of a set of transformer windings 1332b, 1334b and a second voltage clamp 1364. The resulting device circuit 1500 includes a transformer 1530 and a GaN die 1540.

[0085] As in Figure 13ALike transformer 1330, transformer 1530 includes primary winding 1332a, secondary winding 1334a, and optional core 1336a, but does not have a second set of windings for driving the turn-off rectifier.

[0086] GaN die 1540 includes a turn-on rectifier 250 and a pull-down circuit 1560. As in device 1300, the gate-source voltage V of power switch Q1... GS The voltage clamp 1362 is part of the same current loop 1590. Therefore, for example, the effective signal V alternates between +3V and -3V. PRIM_ON A secondary-side voltage is generated, which is rectified by the turn-on rectifier 250 and simultaneously applied to the gate-source of the power switch and across the voltage clamp 1362. Similar to... Figure 13A Circuit 1300, current loop 1590 ensures that pull-down switch Q2 is turned off when power switch Q1 is turned on.

[0087] In addition to the pull-down switch Q2 and voltage clamp 1362, the pull-down circuit 1560 also includes a pull-down resistor R. PD When no ON signal is applied to transformer 1530, pull-down resistor R... PD Discharge the pull-down gate, thereby forcing the pull-down gate-source voltage V. PD_GS The voltage is 0V. This causes the pull-down switch Q2 to turn on and short-circuit the gate and source of the power switch Q1, thus keeping the power switch Q1 in its off state. Similar to the pull-down resistor described earlier, this pull-down resistor R... PD It has a 1kΩ resistor and is generated as a 2DEG channel within the GaN die. Other resistors can be used to change the rate of pull-down gate discharge.

[0088] Device 1500 shares many of the advantages of device 1300, such as providing safe and reliable turn-off of power switch Q1, avoiding shoot-through, duty cycle-related power consumption, and using a single on-current loop to ensure that switches Q1 and Q2 are in their intended state. Device 1500 is also relatively simple, but unlike other circuits described herein that actively turn off power switches by applying a turn-off signal, it does not support dynamic control of the turn-off speed.

[0089] Figure 16 It shows the corresponding Figure 15 The voltage and current of the device 1500. These waveforms are related to... Figure 14 The waveforms are basically the same, except that there is no voltage or current waveform corresponding to the turn-off signal. The turn-off transition of power switch Q1 at time t1 is determined by the rate of discharge of the pull-down gate, which is determined by the pull-down resistor R. PDThe resistance is determined. For example, the switching of power switch Q1 at time t2 is determined by the switch signal V. PRIM_ON The transmitted energy is determined. Therefore, the switching can be altered by changing the voltage of the signal, the current of the signal, the jump pulse, etc.

[0090] Transformer-based switching devices without a pull-down switch

[0091] The transformer-based switching devices described in the previous sections used pull-down switches to actively discharge the power switch control terminal (gate). The following transformer-based switching device controls the conduction of power switch Q1 without a pull-down switch.

[0092] Figure 17 A transformer-driven switching device 1700 is shown, comprising a transformer 1730 and a GaN die 1740. The GaN die 1740 includes a turn-on diode D1 and a turn-off voltage clamp 1762. Figure 13B As shown, the switching diode D1 can be a gate-controlled diode used in the previously described rectifier, and the voltage clamp 1762 can include several gate-controlled diodes stacked together. The switching diode D1 provides half-wave rectification of the AC signal output from the secondary winding 1734a. A capacitor C1 is coupled between the gate and source of the power switch Q1 to smooth the half-wave rectified voltage and maintain a steady-state current to the gate of the power switch during its on-interval. The capacitor C1 can have a capacitance of 300pF or other values ​​and can be a metal-insulator-metal capacitor integrated within the GaN die 1740.

[0093] The primary side signal V used in this device 1700 PRIM_ON V PRIM_OFF It can be single-ended, allowing a single half-bridge to drive each of these signals, rather than the full bridge typically used to generate differential signals as described earlier. Small DC blocking capacitors C2 and C3 (e.g., 100pF) can be included in the transformer 1730 to prevent current drift to excessive values. When the transformer 1730 is a coreless transformer implemented in a silicon die, these capacitors C2 and C3 can be monolithically integrated with the transformer 1730 in the same silicon die.

[0094] The voltage clamp 1762 is used as a diode, but requires a larger bias voltage than diode D1 to ensure operation is not affected during the turn-on interval. In this example, the voltage clamp 1762 comprises three gate-controlled diodes stacked in series, providing a forward bias voltage of 4 × 1.2V = 4.8V. When the turn-off signal V... PRIM_OFFWhen applied to transformer 1730, the off voltage clamp 1762 acts as a half-wave rectifier, and only positive current flows from the gate of the power switch to transformer 1730, thereby discharging the gate of the power switch.

[0095] Figure 17 Device 1700 has the advantage of using a relatively simple circuit and requiring only a single-ended drive signal. Diode D1, voltage clamp 1762, capacitor C1, and power switch Q1 can be monolithically integrated into the same GaN die 1740, providing cost and size advantages, reducing parasitic inductance, and minimizing the length of the current loop. If zero volts is used to turn off power switch Q1, no current is needed to maintain the off state of power switch Q1, and a negative turn-off voltage may be desired to provide a faster turn-off transition, as in device 1700 shown. The disadvantage of device 1700 is that the gate and source of power switch Q1 are no longer directly shorted together, and only when the turn-off signal V... PRIM_OFF When active, the gate voltage is clamped to zero or negative, which may prevent the power switch Q1 from being safely kept off during no-power and startup conditions.

[0096] Figure 18 It shows the corresponding Figure 17 The voltage and current waveforms of device 1700. During the on-interval, the on signal V... PRIM_ON After being boosted by transformer 1730 and rectified by switching diode D1, the voltage alternates between 0V and 3V to provide a turn-on voltage V to the gate of power switch Q1. GS and current I GS During the turn-off interval, the turn-off signal V PRIM_OFF After being boosted by transformer 1730 and rectified by voltage clamp (stacked diodes) 1762, the voltage alternates between 0V and 3V, providing a turn-off voltage of approximately -1.2V. GS This negative voltage is used to make the gate of the power switch discharge more quickly, thereby accelerating the turn-off transition and more robustly keeping the power switch Q1 in its off state.

[0097] Figure 19 A transformer-driven switching device 1900, which also lacks a pull-down switch, is shown. For this device, the primary side is connected by a signal V. PRIM_ON and primary side turn-off signal V PRIM_OFF Differential driving means that a signal generator utilizing a full-bridge can be used. (For illustration, such a signal generator is not shown.) Due to the differentially applied input signal, no DC blocking capacitor is needed within the transformer (e.g., silicon die) 1930. Otherwise, apart from the turns ratio of the two pairs of windings 1932a, 1934a, 1932b, 1934b, transformer 1930 and... Figure 17 The transformer 1730 is essentially the same. These turns ratios can be, for example, 1:2, or can be selected to match a specific rectifier voltage drop and the turn-on threshold voltage of the power switch Q1.

[0098] The GaN die includes a turn-on rectifier 250, a turn-off rectifier 1952, and a power switch Q1. The turn-on rectifier 250 provides full-wave rectification, eliminating the need for a smoothing capacitor at the gate of the power switch. The turn-off rectifier 1952 includes two stacked gated diodes in each branch, preventing unintentional activation of the turn-off rectifier 1952 during the turn-on interval of the power switch Q1. This allows a negative voltage to be applied to the gate of the power switch during the turn-off interval (which discharges more quickly at the gate during the turn-off transition and keeps the power switch Q1 more firmly in its off state).

[0099] Figure 20 It shows the corresponding Figure 19 The voltage and current waveforms of device 1900. During the switching interval, the primary side AC switching voltage V... PRIM_ON It alternates between +3V and -3V. At this voltage V PRIM_ON For example, after being boosted to twice the voltage by transformer 1930 and rectified by turn-on rectifier 250 (which causes a voltage drop of approximately 2.4V), the gate-source voltage V... GS =3.6V and current I GS It is supplied to power switch Q1 and turns it on. During the turn-off interval, the primary side AC turn-off voltage V PRIM_OFF It alternates between +3V and -3V. At this voltage V PRIM_OFF For example, after being boosted to 2.6 times by transformer 1930 and rectified by turn-off rectifier 1952, the gate-source voltage V is approximately -3V. GS An application is made to power switch Q1, thereby keeping power switch Q1 off. Device 1900 provides superior performance compared to... Figure 17 The advantage of device 1700 is that the shutdown switching of device 1900 is much faster.

[0100] Figure 21 A transformer-driven switching device 2100 is shown, except that the turn-off is achieved by a pull-down resistor R instead of a voltage clamp 1762. PD In addition to the provided features, the transformer-driven switching device 2100 is similar to... Figure 17 The transformer-driven switching device simplifies the circuitry of the GaN die 2140. In contrast, the transformer (e.g., a silicon die) 2130 has only one set of windings 1732a and 1734a. The primary side on-state voltage V... PRIM_ON It can be single-ended. During the on-interval, the primary side is connected to the signal V.PRIM_ON Alternating between two values ​​(e.g., 0 and 3V), the resulting secondary-side voltage and current are rectified by diode D1. Because diode D1 only provides half-wave rectification, capacitor C1 (shown as 500pF) may be useful to smooth the voltage and current supplied to the gate of the power switch during the turn-on interval. During the turn-off interval, pull-down resistor R... PD Discharging the gate of the power switch passively turns off the power switch Q1. Device 2100 has the advantages of simple circuitry and minimal design, but it has the disadvantage of a relatively slow turn-off transition. However, applications that do not require a fast turn-off transition, such as certain synchronous rectifier bridges, may benefit from the simple circuitry provided by device 2100.

[0101] Figure 22 It shows the corresponding Figure 21 The voltage and current waveforms of device 2100. These waveforms are similar to... Figure 20 The waveform, but a notable exception is the absence of a turn-off signal (V). PRIM_OFF I PRIM_OFF I SEC_OFF Similarly, the gate-source voltage V of power switch Q1 can be observed. GS It decays more slowly. (Smaller pull-down resistor R) PD This will result in faster shutdown transitions, but may come at the cost of slower turn-on transitions and higher power consumption during the turn-on interval.

[0102] System-in-Package (SiP) for Transformer-Based Switching Devices

[0103] Any transformer-based switching device described herein can be advantageously integrated into the same package. Figure 23 It shows, for example, corresponding to Figure 2A Examples of such system-in-package 2300 for device 200. A coreless transformer 230, preferably implemented within a silicon die, has an input with leads 2302 bonded to the package 2300. The output of the coreless transformer 230 is provided to a GaN die 240, which includes a power switch Q1 and drive circuitry, for example, a transformer-driven switching device 200 as described herein. The source terminal of the power switch Q1 within the GaN die 240 is bonded to a lead frame of the package 2300, which in turn is connected to the leads of the package 2300. The drain of the power switch Q1 is, for example, wire bonded. (wire bond) The leads are connected to the package 2300. The SiP provides an integrated device that includes a power switch Q1 and any necessary level shifting, isolation, driver power, etc.

[0104] Transformer-based bidirectional switching device

[0105] The previous description focused on transformer-based switching devices, each comprising discrete GaN switches. Advantageously, the described drive circuitry can be applied to other topologies, other power switch types, and multiple switches within a single package. Figure 24 An example is shown in which two GaN power switches are integrated into a single GaN die, and a single transformer chip (e.g., a coreless transformer implemented in a silicon die) provides drive signals to the GaN die. The power switches in the GaN die can be controlled individually or together to achieve ideal switching, i.e., a bidirectional switch capable of blocking current in both directions. The illustrated device 2400 reduces the high cost and complex drive circuitry typically associated with driving bidirectional switches, particularly bidirectional GaN switches. Figure 24 The transformer-driven bidirectional switching device 2400 is used with Figure 15 The circuit is similar to that of device 1500. However, it should be understood that other transformer-driven switching devices described previously can be arranged similarly to obtain bidirectional switching.

[0106] The bidirectional switching device 2400 includes having a connection with Figure 15 The device 2400 comprises a first (lower) portion of the same circuit as the first portion. The lower portion includes first transformer windings 2402a and 2404a, an optional first core 2406a, a first turn-on rectifier 250, a first pull-down circuit 1560, and a first power switch Q1. The device 2400 also includes a second (upper) portion having circuitry mirrored that of the first portion. The second portion includes second transformer windings 2403a and 2405a, an optional second core 2407a, a second turn-on rectifier 251, a second pull-down circuit 1561, and another power switch Q3. The drain terminals D1 and D3 of power switches Q1 and Q3 are coupled together. Source terminals S1 and S3 of power switches Q1 and Q3 are provided for external connection to a GaN bidirectional switch die 2440. As shown, the bidirectional switch device 2400 is distributed across a silicon die including a transformer 2430 and a GaN die 2440 including power switches Q1 and Q3 and associated circuitry for controlling the power switches.

[0107] Transformer-based switching device with normally open power switch

[0108] Previous examples of transformer-based switching devices include normally-off power switches. The circuit topologies previously described can be adapted to control normally-open power switches. Figure 25 A transformer-driven switching device 2500, including a normally open power switch Q1, is shown. Device 2500 includes a transformer 2530 and a GaN die 2540. The circuitry of device 2500 is... Figure 15The circuit of device 1500 is in an inverted form, but includes an additional resistor R2 to couple the source of power switch Q1 to its gate. Figure 15 The power switch Q1 in device 2500 is opposite; when a valid signal is applied to transformer 2530, power switch Q1 of device 2500 is turned off, otherwise it is turned on.

[0109] By default, the normally open pull-down switch Q2 is turned on, thus short-circuiting the source and gate of power switch Q1. Its gate-source voltage V... GS When the value is 0, the normally open power switch is turned on in this default state. After the turn-off interval, the pull-down resistor R1 makes the voltage across the pull-down gate and pull-down source equal, thereby forcing the pull-down switch Q2 and the power switch Q1 back to their default (on) state.

[0110] When the effective shutdown signal V is applied PRIM_OFF For example, when a signal alternating between +3V and -3V is applied to transformer 2530, rectifier 2550 rectifies the corresponding secondary-side signal and supplies current to current loop 2590. The voltage supplied by rectifier 2550 to current loop 2590 is the gate-source voltage V of power switch Q1. GS This is allocated between the voltage clamp 2562 and the voltage clamp 2562. The voltage clamp 2562 can be constructed using a gated diode, which in this example provides an 8V forward bias threshold (clamping voltage V). CL The additional resistor R2 provides a parallel current path to the source-gate path. The parallel path of resistor R2 maintains a continuous current during the turn-off interval to keep the negative gate-source voltage Vo. GS This keeps the power switch Q1 off.

[0111] Figure 26 It shows the corresponding Figure 25 The voltage and current waveforms of the transformer-driven power switching device 2500 are shown. During the off-state interval of power switch Q1, the primary side disconnects the AC signal V. PRIM_OFF The voltage alternates between +3V and -3V. This voltage is boosted by transformer 2530 and the voltage drop is reduced by diodes within rectifier 2550. The voltage obtained from the rectifier 2550 output is applied to the gate-source terminals of power switch Q1 and across voltage clamp 2562 and has an amplitude of approximately 16V. Voltage clamp 2562 has a clamping voltage V when forward biased. CL =8V, as shown in the figure, resulting in a pull-down gate-source voltage V of -8V during the turn-off interval. PD_GS This keeps the pull-down switch Q2 in its off state. As shown in the figure, the residual voltage output from the rectifier 2550 provides a -8V gate-source voltage V to the power switch Q1.GS This keeps the power switch Q1 in its off state. During the off interval, the current I... GS The current flows out from rectifier 2550. During the on-off interval, the primary side turns off the AC signal V. PRIM_OFF It is ineffective, for example, if it is not driven or driven at 0V or some other DC voltage. No significant energy is transferred across the transformer 2530, and no current I is supplied. GS Turn off power switch Q1. Pull down gate-source voltage V. PD_GS and the gate-source voltage V of the power switch GS When the voltage rises to 0V, the power switch Q1 is normally open by default.

[0112] Figure 27 A transformer-driven switching device 2700 is shown, comprising two normally open power switches Q1 and Q3 arranged as bidirectional switches. The circuitry for driving these switches is similar to that used for driving... Figure 25 The circuit of device 2500. However, this circuit is compatible with... Figure 24 The bidirectional switch based on the normally closed power switch described in the paper is replicated and inverted in the same way to drive the power switch Q3.

[0113] Transformer-based switching device with non-GaN power switch

[0114] Previous examples of transformer-based switching devices focused on devices incorporating GaN transistors as power switches. However, the circuit topology can be similarly applied to other power switches, including power MOSFETs. Such MOSFET-based devices preferably have different partitions between the dies. Notably, the rectifier diode, pull-down switch, and / or associated drive circuitry can be integrated with the transformer on the same die. As in the previously described devices, both active and passive gate pull-down can be used with power MOSFETs. However, when using a coreless transformer, the switching speed may be slower than with conventional MOSFET drivers. This can be improved through various techniques. For example, a stronger current-handling coil with reduced coil impedance may result in faster switching transitions. Cored transformers can be used, which provide improved coupling, resulting in higher power delivery and faster switching transitions, but may make integration difficult. Using a full-bridge differential drive signal to turn on the power MOSFET results in higher power delivery and reduced switching transition time compared to a single-ended drive signal. Active pull-down of the power MOSFET gate typically provides a faster turn-off transition, and in some applications, it may be necessary to achieve a turn-off speed acceptable to the power MOSFET.

[0115] Figure 28A transformer-based switching device 2800 is shown, which includes a MOSFET 2860 as a power switch Q1. The example shown uses a similar... Figure 19 The circuitry of device 1900 is physically divided. Device 2800 includes a transformer and a rectifier die 2840. The transformer 1930, the turn-on rectifier 250, and the turn-off rectifier 1952 are monolithically integrated within the die 2840. The die 2840 may be a silicon die, a silicon carbide die, or made of some other semiconductor material. As shown, a power MOSFET 2860 is integrated on a separate die.

[0116] Because transformer 1930, switching rectifier 250, and switching rectifier 1952 are similar to those concerning… Figure 19 The corresponding components are described, so they will not be fully described. However, note that the turn-off rectifier 1952 has twice as many diodes as the turn-on rectifier 250. This prevents the diodes within the turn-off rectifier 1952 from being forward biased (activated) during the turn-on interval, during which a positive voltage is driven across the gate and Kelvin source (KS) terminals of the MOSFET 2860. In particular, the turn-off rectifier 1952 should have a sufficiently high threshold voltage so that it is not forward biased to a level as high as the maximum expected gate-source voltage V of the MOSFET 2860. GS Any voltage.

[0117] Because the gate of a power MOSFET is purely capacitive and does not require current to be supplied throughout the entire on or off interval, the energy transfer cycle can be limited to the time interval corresponding to the switching transition. Therefore, as... Figure 29 As shown, the drive signal on and the drive signal off may be discontinuous.

[0118] Figure 29 It shows the corresponding Figure 28 The voltage and current waveforms of the transformer-based switching device 2800 are shown. Starting from time t0, the activation signal V is... PRIM_ON Make the on signal V PRIM_ON The voltage alternates between +10V and -10V until time t0a. During this interval, a primary-side current I is supplied. PRIM_ON and secondary side current I SEC_ON And after rectification by the turn-on rectifier 250, a positive current I is supplied to the gate of the MOSFET 2860. GS Once the gate-source voltage V GS When the turn-on threshold is reached, for example, 5V for the power MOSFET 2860, the power MOSFET 2860 will turn on. After the power MOSFET 2860 is turned on, the turn-on signal V...PRIM_ON It can be deactivated, for example, by turning on the signal V. PRIM_ON It is set to 0V or some other DC voltage.

[0119] For the turn-off transition at time t1, the activation turn-off signal V is... PRIM_OFF This causes it to alternate between +10V and -10V until time t1a. During this interval, a primary-side current I is provided. PRIM_OFF and secondary side current I SEC_OFF And after being rectified by the turn-on rectifier 250, it draws current from the gate of the MOSFET 2860, that is, provides a negative current I. GS This draw current discharges the MOSFET gate (G) and reduces the gate-source voltage V. GS This turns off the power MOSFET 2860. After the power MOSFET 2860 is turned off, the on signal V is activated. PRIM_ON It can be deactivated, for example, by being set to 0V or some other DC voltage. The turn-on and turn-off speeds depend in part on the gate-source capacitance of the MOSFET 2860, which is 2nF for the embodiment shown.

[0120] The circuit topology described for a transformer-based switching device using GaN power switches can be compared with... Figure 28 The same method shown is similarly modified to use a MOSFET power switch.

[0121] Switching speed

[0122] The transformer-driven switching device described in this paper offers considerable flexibility in terms of switching speed, which provides advantages compared to conventional power switch driving technologies. For applications based on on-signal such as V... PRIM_ON The active switching device can dynamically control the switching speed by changing the voltage and / or current provided by the switching signal. Alternatively, the switching speed can be slowed down by skipping pulses within the switching signal, which may be simpler because, unlike changing the analog power signal, this only requires changing the digital control in the signal generator. Similarly, the duty cycle or frequency of the switching signal can be changed to control the switching speed. Using a switching signal such as V... PRIM_OFF The active turn-off device can similarly alter its turn-off transition speed by changing the voltage and / or current of the turn-off signal and / or by using pulse skipping or changing the duty cycle or frequency of the turn-off signal. For ideal power switches such as GaN that have a high drive current when turned on but may subsequently maintain a low current, pulse skipping can be used to achieve a lower holding current.

[0123] Although this disclosure is not limited thereto, the examples numbered below illustrate one or more aspects of this disclosure.

[0124] Example 1. A transformer-driven switching device includes: a power transistor, a first transformer, and a first rectifier. The power transistor has a first load terminal and a second load terminal, and a control terminal for controlling conduction between the first load terminal and the second load terminal. The first transformer includes a first primary winding and a second secondary winding. The first rectifier is coupled between the first secondary winding and the control terminal of the power transistor. The first secondary winding and the first rectifier are configured to transmit required energy and control timing for driving the control terminal based on an input waveform coupled to the first primary winding, so as to control conduction between the first load terminal and the second load terminal. The input waveform includes turn-on intervals having a plurality of high-frequency pulses, the first transformer and the first rectifier converting the plurality of high-frequency pulses into a single turn-on control pulse provided to the control terminal to turn on the power transistor, and the input waveform includes a DC level during the turn-off interval of the power transistor.

[0125] Example 2. A transformer-driven switching device according to Example 1, wherein the power transistor, the first transformer and the first rectifier are integrated in the same package.

[0126] Example 3. A transformer-driven switching device according to Example 1, wherein the first transformer is a coreless transformer without a magnetic core.

[0127] Example 4. A transformer-driven switching device according to Example 3, wherein the first transformer is integrated in a silicon die.

[0128] Example 5. A transformer-driven switching device according to Example 4, wherein the first rectifier and the first transformer are integrated in the same silicon die.

[0129] Example 6. A transformer-driven switching device according to Example 1, wherein the frequency of the high-frequency pulse varies during the turn-on interval in order to control the turn-on speed of the power transistor.

[0130] Example 7. A transformer-driven switching device according to Example 1, wherein the power transistor is a normally-off high electron mobility transistor (HEMT), the first load terminal is the drain, the second load terminal is the source, and the control terminal is the gate.

[0131] Example 8. A transformer-driven switching device according to Example 7, wherein the normally-off HEMT is a gallium nitride (GaN) gate-injected transistor.

[0132] Example 9. The transformer-driven switching device according to Example 7 further includes a pull-down transistor configured to short-circuit the gate to the source during the off-state interval of the power transistor.

[0133] Example 10. A transformer-driven switching device according to Example 9, wherein the pull-down transistor is a normally open HEMT including a pull-down drain, a pull-down source, and a pull-down gate. The transformer-driven switching device further includes a pull-down control circuit connected between the pull-down gate and the pull-down source, and configured to autonomously apply a negative pull-down turn-off voltage relative to the pull-down source to the pull-down gate when the power transistor is turned on, and to autonomously discharge the pull-down turn-off voltage during the turn-off interval of the power transistor.

[0134] Example 11. A transformer-driven switching device according to Example 10, wherein the pull-down control circuitry includes a voltage clamp and a resistor. The voltage clamp is coupled between the pull-down source and the pull-down gate and is configured to apply the pull-down turn-off voltage relative to the pull-down source to the pull-down gate when the power transistor is turned on. The resistor is coupled between the pull-down gate and the pull-down source and is configured to provide self-discharge of the pull-down turn-off voltage when the power transistor is turned off.

[0135] Example 12. A transformer-driven switching device according to Example 11, wherein the voltage clamp includes one or more HEMTs configured as diodes, the resistor includes a two-dimensional electron gas (2DEG) HEMT region, and the first rectifier includes one or more HEMTs configured as diodes. The first rectifier, the pull-down transistor, and the pull-down control circuit are monolithically integrated with the normally off HEMT on the same die.

[0136] Example 13. The transformer-driven switching device according to Example 9 further includes a second transformer, the second transformer including a second primary winding and a second secondary winding. The pull-down transistor is a normally-off HEMT including a pull-down source, a pull-down drain, and a pull-down gate. The transformer-driven switching device further includes a second rectifier coupled between the second secondary winding and the pull-down gate. The second rectifier is configured to convert a plurality of high-frequency pulses received from the second secondary winding into a single turn-on pulse applied to the pull-down gate.

[0137] Example 14. A transformer-driven switching device according to Example 13, wherein the pull-down drain is coupled to the gate of the power transistor, the pull-down gate is coupled to the source of the power transistor, and the second rectifier is coupled across the pull-down gate and the pull-down source.

[0138] Example 15. A transformer-driven switching device according to Example 14, wherein the frequency and / or amplitude of a high-frequency turn-off waveform applied to the second primary winding controls the turn-on speed of the pull-down transistor and the turn-off speed of the power transistor.

[0139] Example 16. A transformer-driven switching device according to Example 13, wherein the first rectifier, the second rectifier, and the pull-down transistor are monolithically integrated on the same die as the power transistor.

[0140] Example 17. The transformer-driven switching device according to Example 7 further includes a second transformer and a second rectifier. The second transformer includes a second primary winding and a second secondary winding. The second rectifier is coupled to the second secondary winding and is coupled across the gate and source of the power transistor such that a negative turn-off voltage is provided between the gate and the source in response to an alternating current (AC) signal received from the second secondary winding.

[0141] Example 18. A transformer-driven switching device according to Example 17, wherein the first rectifier and the second rectifier are monolithically integrated with the power transistor in the same die.

[0142] Example 19. A transformer-driven power switching system includes a power transistor, a first transformer, a rectifier, and control logic circuitry. The power transistor includes a first load terminal, a second load terminal, and a control terminal for controlling conduction between the first load terminal and the second load terminal. The first transformer includes a first primary winding and a second secondary winding. The rectifier is coupled between the transformer and the control terminal. The control logic circuitry includes a control input and a frequency generator, wherein the frequency generator is configured to provide a high-frequency on signal to the first primary winding in response to receiving an on level at the control input, and to provide a direct current (DC) signal to the first primary winding in response to receiving an off level at the control input.

[0143] Example 20. A transformer-driven power switching system according to Example 19, wherein the control logic circuit and the first transformer are monolithically integrated in the same die.

[0144] As used herein, the terms “having,” “containing,” “including,” “comprising,” etc., are open-ended terms that indicate the presence of a stated element or feature but do not exclude additional elements or features. Unless the context clearly indicates otherwise, the articles “a,” “an,” and “the” are intended to include both plural and singular forms.

[0145] It should be understood that, unless otherwise specifically stated, the features of the various embodiments described herein can be combined with each other.

[0146] While specific embodiments have been illustrated and described herein, those skilled in the art will understand that various alternative and / or equivalent implementations may be substituted for the specific embodiments shown and described without departing from the scope of the invention. This application is intended to cover any adaptations or variations of the specific embodiments discussed herein. Therefore, the invention is intended to be limited only by its claims and their equivalents.

Claims

1. A transformer-driven switching device, comprising: A power transistor, wherein the power transistor is implemented as a normally-off high electron mobility transistor, i.e., a normally-off HEMT, and includes a drain, a source, and a gate for controlling the conduction between the drain and the source; The first transformer includes a first primary winding and a first secondary winding; A first rectifier is coupled between the first primary winding and the gate. A normally open pull-down transistor, the pull-down transistor being configured to short-circuit the gate of the power transistor to the source during the off-interval of the power transistor; as well as A voltage clamp having a first terminal connected to the source of the power transistor and the pull-down source of the pull-down transistor; and a second terminal connected to the pull-down gate of the first rectifier and the pull-down transistor. The primary winding and the first rectifier are configured to transmit the energy required to drive the gate and control timing based on an input waveform coupled to the primary winding, in order to control the conduction between the drain and the source. The input waveform includes turn-on intervals with a plurality of high-frequency pulses. The first transformer and the first rectifier convert the plurality of high-frequency pulses into a single turn-on control pulse provided to the gate to turn on the power transistor. The input waveform includes a DC voltage level during the turn-off interval of the power transistor. The voltage clamp is configured to drive a negative voltage to the pull-down gate of the pull-down transistor when the power transistor is turned on, thereby forcing the pull-down transistor to a turn-off state.

2. The transformer-driven switching device according to claim 1, wherein, The power transistor, the first transformer, and the first rectifier are integrated into the same package.

3. The transformer-driven switching device according to claim 1, wherein, The first transformer is a coreless transformer without a magnetic core.

4. The transformer-driven switching device according to claim 3, wherein, The first transformer is integrated into a silicon die.

5. The transformer-driven switching device according to claim 4, wherein, The first rectifier and the first transformer are integrated in the same silicon die.

6. The transformer-driven switching device according to claim 1, wherein, The frequency of the high-frequency pulse varies during the on-time interval in order to control the on-time of the power transistor.

7. The transformer-driven switching device according to claim 1, wherein, The normally-off HEMT is a gate-injected transistor based on gallium nitride (GaN).

8. The transformer-driven switching device according to claim 1, wherein, The pull-down transistor is a normally open high electron mobility transistor (HEMT) and also includes a pull-down drain.

9. The transformer-driven switching device according to claim 1, further comprising: A resistor coupled between the pull-down gate and the pull-down source of the pull-down transistor.

10. The transformer-driven switching device according to claim 9, wherein, The voltage clamp includes one or more HEMTs configured as diodes, the resistor includes a two-dimensional electron gas (2DEG) region, the first rectifier includes one or more HEMTs configured as diodes, and the first rectifier and the pull-down transistor and the normally off HEMT are monolithically integrated on the same die.

11. A transformer-driven power switching system, comprising: A transformer-driven switching device according to any one of claims 1 to 10; as well as A control logic circuit includes a control input and a frequency generator, wherein the frequency generator is configured to provide a high-frequency signal to the first primary winding in response to receiving an on level at the control input, and to provide a DC signal to the first primary winding in response to receiving an off level at the control input.

12. The transformer-driven power switching system according to claim 11, wherein, The control logic circuit and the first transformer are monolithically integrated into the same die.

Images