Regenerative gate drive circuit for power MOSFET

1. Field of the Invention

This invention relates generally to switched mode power converters and in particular to a regenerative gate drive circuit for voltage control of MOSFETs (Metal Oxide Silicon Field Effect Switching device) and IGBT (Insulated Gate Bipolar Switching device) having means for recovering energy stored in the gate capacitance and means for speeding up gate voltage transitions and reducing output impedance.

2. Description of the Related Art

Losses in the gate drive circuits of switched mode power converters operating in the hundreds of kilohertz and megahertz regions significantly affect their efficiency and complicate thermal design.

As conventional prior art gate drive circuit is shown in FIG. 1 comprising a totem pole pair of field effect switching devices Q 1 and Q 2 and a gate resistor, R g . This topology suffers from excessive power dissipation particularly at high frequency. The three major sources of power dissipation are as follows: 1. capacitive energy losses, 2. cross conduction power losses, and 3. linear operation losses.

1. Capacitive energy losses occur during turn-on and turn-off. During turn-on, the gate energy (½C gs V dd 2 ) is stored within the equivalent gate capacitance and the same amount of energy is dissipated in the total series resistance (gate driver resistance R g and internal gate resistance of the switch Q 2 ). During turn-off, the same energy stored in the gate (½C gs V dd 2 ) is dissipated in the total series resistance. Therefore, the total dissipated power is:

P capacitive =ƒC gs V dd 2

where

ƒis switching frequency

C gs is gate-to-source capacitance

V dd is supply voltage

The power is dissipated within the gate circuitry R g and the internal MOSFET gate resistance. While the energy is independent of the charge and discharge path resistance, this resistance determines the switching times and the current rating of the gate-drive switches.

2. Cross conduction power losses occur when in order to generate narrow pulses, one of the totem pole resistors is turned on before the other one is turned completely off. This mode of operation results in cross conduction losses because both gate drive switching devices conduct simultaneously during switching transitions.

3. Linear operation losses are caused by an overlap of voltage and current across gate drive switching devices during switching transitions.

Another problem of the conventional circuit of FIG. 1 is that the parasitic inductance in series with the gate limits the rate of change in the gate current, and thereby increases the switching time. During switching transients, current flowing through the ‘off’ switch Miller capacitance will go through the gate capacitance as well, turning the device back on, unless an alternative low impedance path is available. In the resistive circuit, this impedance from Z g with the switch Q 2 in the on state, where Z g is the total signal and return path impedance from the power supply to the MOSFET's gate.

In a paper by W. A. Tobisz et al, entitled “Zero-Voltage-switched quasi-resonant buck and flyback converters—experimental results at 10 MHz”, Proceedings of IEEE PESC '87 Conference, 1987, pp 404-413, a quasi-resonant gate-drive offers reduced losses at turn-on of the power MOSFET. This circuit does not clamp the gate voltage and may cause an over-voltage, and the gate energy is dissipated within the switch at turn-off.

Circuits described in T. Chen, et al., “A resonant MOSFET gate-driver with complete energy recovery” Proceedings of the 3rd IEEE Power Electronics and Motion Control Conference (IPEMC) 2000, Vol. 1, p. 402-406, and I. D. de Vries, “A resonant power MOSFET/IGBT gate driver”, Applied Power Electronics Conference and Exposition, APEC 2002, vol. 1, pp. 179-185, recover gate energy by reversing the voltage across the resonant inductor once the energy has been transferred to it from the gate capacitance. Topology in Y. Chen, et al. clamps the gate voltage to the source through additional pair of semiconductor devices, but does not provide protection against cross conduction and has low output impedance only during conduction of the clamp diodes. While the gate drive circuit in I. D. de Vries et al, prevents cross conduction, it has high output impedance that makes the MOSFET gate susceptible to false triggering.

In a paper by D. Maksimovic entitled “A MOS gate drive with resonant transitions”, Record of the 22nd IEEE Annual Power Electronics Specialists Conference (PESC), 1991, pp. 527-532, it describes a MOS gate drive circuit that clamps gate voltage to the source and prevents cross conduction during normal operation (excluding operation in the presence of Electromagnetic Pulse of EMP). However, it requires placement of gate drive switching devices close to the MOSFET gate to minimize wiring inductance.

U.S. Pat. No. 5,264,736 issued Nov. 23, 1993 to B. Jacobson discloses a high frequency resonant gate drive for a Power MOSFET. This regenerative topology enables partial energy recovery and protects against cross-conduction. It also provides low output impedance because currents circulate in clamp diodes during the time interval between the gate voltage transitions. However, energy recovery takes place only during the narrow time interval prior to the turn off of the main switching device when voltage across the resonant inductor exceeds the bias supply.

U.S. Pat. No. 6,208,535 issued Mar. 27, 2001 to David Parks discloses a resonant gate driver that provides a resonant switching for a field effect transistor (FET) and associated circuits such as power supplies with synchronous output rectifiers. The resonant switching for a junction FET invokes bipolar mode operation with a diode clamping of the gate, which results in bipolar mode operation without a separate bias power supply. However, this resonant gate driver does not provide energy recovery for efficiency improvement.

In a paper by Patrick Dwane, et al. entitled “An Assessment of Resonant Gate Drive Techniques for use in modem Low Power DC-DC converters” IEEE, Jan. 2005, pp. 1572-1580, various gate drive topologies are reviewed and described including several of the above-identified prior art references.

Accordingly, it is therefore an object of this invention to provide a regenerative gate drive circuit that improves noise immunity, speeds up switching transition times, and reduces power consumption.

It is another object of this invention to provide a tuned network to speed-up voltage transitions and reduce output impedance.

It is a further object of this invention to reduce losses in the gate drive circuits of a switched-mode power converter.

These and other objects are further accomplished by providing a regenerative gate drive circuit comprising a positive power source coupled to a first switching device, a negative power source coupled to a second switching device, a first inductor network comprises a first winding and a second winding on each side of a center tap, said center tap being coupled to a gate input of a power MOSFET, a second inductor network comprises a first winding in series with a first blocking capacitor and a second winding in series with a second blocking capacitor on each side of a center tap, the center tap being coupled to the gate input of a power MOSFET and to the center tap of the first inductor network, the first switching device, coupled between the positive power source and the first winding of the first inductor network in parallel with the first blocking capacitor in series with the first winding of the second inductor network, provides a path for a first current to flow through the first switching device and the first winding of the first inductor network in parallel with the first winding of the second inductor which is in series with the first blocking capacitor, the second switching device, coupled between the negative power source and the second winding of the first inductor network in parallel with the second winding of the second inductor network, the second winding being in series with a second blocking capacitor, provides a path for resonant discharging of the power MOSFET, a first input signal is coupled to a gate of the first switching device for controlling the first switching device, a second input signal is coupled to a gate of the second switching device for controlling the second switching device, means coupled to the first switching device and the second switching device for limiting the voltage provided at the gate of the power MOSFET and for conducting inductive current, and means coupled between the positive power source and the second switching device and between the negative power source and the first switching device for limiting the voltage at the gate of the power MOSFET and for conducting inductive current.

Each of the first switching device and the second switching device comprises a MOSFET. The first input signal is coupled to an input gate of the first MOSFET, and second input signal is coupled to an input gate of the second MOSFET. The first winding and the second winding of the first inductor network and the second inductor network comprises an equal number of turns. The limiting means comprises a first diode coupled between the positive power source and the second switching device and a second diode coupled between the negative power source and the first switching device. The second inductor network provides reduced output impedance thereby improving noise immunity. The first inductor network and the second inductor network provide a high impedance between the first switching device and the second switching device when the switching devices are both conducting simultaneously. The circuit comprises a dead time interval when the first input signal and the second input signal are at a level that turns-off the first switching device and the second switching device thereby enabling energy recovery. Control of the dead time interval is provided by means for sensing any one of an output voltage or output current of the power MOSFET, or power across the power MOSFET or temperature at a junction of the power MOSFET. These objects are further accomplished by a method for providing a regenerative gate drive circuit comprising the steps of providing a positive power source coupled to a first switching device, providing a negative power source coupled to a second switching device, providing a first inductor network having a first winding and a second winding on each side of a center tap, said center tap being coupled to a gate input of a power MOSFET, providing a second inductor network having a first winding in series with a first blocking capacitor and a second winding in series with a second blocking capacitor on each side of a center tap, the center tap being coupled to the gate input of a power MOSFET and to the center of tap of the first inductor network, providing a path for a first current to flow through the first switching device and the first winding of the first inductor network in parallel with the first winding of the second inductor which is in series with the first blocking capacitor, the first switching device being coupled between the positive power source and the first winding of the first inductor network in parallel with the first blocking capacitor in series with the first winding of the second inductor network, providing a path for resonant discharging of the power MOSFET through the second switching device coupled between the negative power source and the second winding of the first inductor network in parallel with the second winding of the second inductor network, the second winding being in series with a second blocking capacitor, controlling the first switching device by a first input signal coupled to a gate of the first switching device, controlling the second switching device by a second input signal coupled to a gate of the second switching device, limiting the voltage provided at the gate of the power MOSFET and conducting inductive current by means coupled to first switching device and the second switching device, limiting the voltage provided at the gate of the power MOSFET and conducting inductive current by means coupled between the positive power source and the second switching device and between the negative power source and first switching device.

The method includes the step of providing a MOSFET for each of the first switching device and the second switching device. The method comprises the steps of coupling the first input signal to an input gate of the first MOSFET, and coupling the second input signal to an input gate of the second MOSFET. The method further comprises the step of providing the first winding and the second winding of the first inductor network and the second inductor network with an equal number of turns. The step of limiting the voltage of the gate of the power MOSFET comprises the steps of providing a first diode coupled between the positive power source and the second switching device, and providing a second diode coupled between the negative power source and the first switching device.

Additional objects, features and advantages of the invention will become apparent to those skilled in the art upon consideration of the following detailed description of the preferred embodiments exemplifying the best mode of carrying out the invention as presently perceived.

The appended claims particularly point out and distinctly claim the subject matter of this invention. The various objects, advantages and novel features of this invention will be more fully apparent from a reading of the following detailed description in conjunction with the accompanying drawings in which like reference numerals refer to like parts, and in which:

FIG. 1 is a schematic diagram of a prior art gate drive circuit having a single bias power supply.

FIG. 2 is a schematic diagram of a regenerative gate drive circuit according to the present invention.

FIG. 3 is a graph of typical voltage and current waveforms at the gate of MOSFETS Q 1 and Q 2 and typical waveforms of the current through inductors L 1 , L 2 , L 3 and L 4 of FIG. 2 .

FIG. 4 is a graph of typical current waveforms through the voltage sources V dd and V ss , diodes D 1 and D 2 and switching devices Q 1 and Q 2 .

FIG. 5 is a graph of measured power MOSFET gate voltage generated by the gate drive circuit described in U.S. Pat. No. 5,264,736.

FIG. 6 is a graph of a measured power MOSFET gate voltage generated by the regenerative gate drive circuit of FIG. 2 .

Referring to FIG. 2 , a circuit diagram of a regenerative gate drive circuit 10 is shown embodying the principles of the invention. The regenerative gate drive circuit 10 comprises a pair of MOSFET (metal oxide silicon field effect switching) devices Q 1 , Q 2 connected in a series with coupled inductors L 1 , L 2 between them, the coupled inductors having a center tap 25 which is connected to a gate of a power MOSFET 38 . The coupled inductors L 1 , L 2 comprise an upper half winding and a lower half winding each having an inductance of 300 mH with an inductor coupling factor (k) of 0.92.

Also, connected in parallel with the main coupled inductors L 1 , L 2 are coupled inductors L 3 , L 4 with an inductor coupling factor (k) of 0.92 and blocking capacitators C 1 , C 2 . Blocking capacitor C 1 is connected in series with inductor L 3 and another blocking capacitator C 2 is connected in series with inductor L 4 . The center tap 27 of coupled inductor. L 3 , L 4 connects to the center tap 25 of L 1 and L 2 .

A power or voltage source V dd has a positive terminal connected to a drain terminal of MOSFET Q 1 and has a negative terminal connected to a positive terminal of voltage source V gs . The negative terminal of V ss connects to the source of MOSFET Q 2 . Two clamping diodes D 1 , D 2 are provided wherein diode D 1 is connected from the voltage source V dd to the drain of MOSFET Q 2 , and diode D 2 is connected from the voltage source V ss to the source of MOSFET Q 1 . The diodes D 1 , D 2 clamp the voltages at the gate of the power MOSFET 38 to V dd and V ss respectively which limits the voltage across the drive MOSFET switching devices Q 1 , Q 2 to the sum of the two voltage sources V dd and V ss . In FIG. 2 V dd provides 12.6 volts DC and V ss provides 5 volts DC.

Referring to FIG. 2 , FIG. 3 and FIG. 4 , the operation of FIG. 2 is described by referring to the waveforms shown in FIG. 3 and FIG. 4 . FIG. 3 is a graph of typical voltage and current waveforms at the gates of MOSFETS Q 1 and Q 2 and typical waveforms of the current through coupled inductors L 1 and L 2 and coupled inductors L 3 and L 4 . FIG. 4 is a graph showing typical current waveforms through the voltage sources V ss and V ss , diodes D 1 and D 2 , and MOSFET switching devices Q 1 and Q 2 .

At the beginning of the switching cycle at time t=0, the gate signal V gate Q 2 provided by signal VG 2 for the switching device Q 2 is switched off (low) and it starts turning-off. Switching device Q 1 is not turned-on yet, so the switching cycle begins with a dead time interval 41 when gate signals V gate Q 1 and V gate Q 2 to both switching devices are off. At t=0, output of the gate driver 10 (gate voltage of the power MOSFET 38 ) is negatively biased, diode D 1 is off, and diode D 2 conducts current through inductors L 1 , L 2 , and Q 2 . As Q 2 turns off, its voltage begins to rise and current in the inductor L 2 diverts to the inductor L 4 . At the same time, current in the inductor L 1 starts flowing through the power MOSFET 38 gate, the bias voltage source V ss (thereby regenerating energy), and the diode D 2 . Thus, at the end of the dead time interval 41 (t=50 ns), the power MOSFET 38 gate capacitance is partially discharged, switching devices Q 1 and Q 2 are off, currents in both diodes D 1 and D 2 are close to zero, and both coupled inductors L 1 , L 2 and L 3 , L 4 conduct current.

The dead time interval 41 ends at t≈50 ns when the gate signal V gate Q 1 from VG 1 starts turning on switching device Q 1 . As Q 1 turns on, the voltage is applied to the coupled inductors L 1 , L 2 and L 3 , L 4 and the power MOSFET 38 gate is resonantly charged by the sum of two currents. The source V dd generates the first current that flows through Q 1 and L 1 in parallel with L 3 . The second current is the one that was stored in L 2 and L 4 and it flows through the integral body diode of the switching device Q 2 returning energy to the source V ss . At the end of the positive half wave of the resonant current I gate 46 , the current in the integral body diode of the switching device Q 2 changes direction and starts flowing through its channel while voltage builds up across this switching device Q 2 . While Q 2 conducts current, the voltage source V ss forces current increase in inductors L 2 and L 4 . At the end of the charge-up time interval (t=100 ns), the power MOSFET gate capacitance is almost fully charged, Q 1 is on, Q 2 is off, and inductor currents continue increasing.

Because the resonant gate drive circuit 10 has such non-linear elements as the power MOSFET 38 gate and reverse transfer capacitance, it is exceedingly difficult to obtain a closed form solution for its voltages and currents. However, the following equations derived from a simplified LCR equivalent circuit with square wave excitation provide an assessment of the peak gate drive current I pk and its resonant frequency (f) during this time interval.

I peak =( V dd −0.5( V dd +V ss )−V Gate(0) )/2π f ( L−M )   (1)

f=✓( 1/( L−M ) C eq −R/ 2( L−M ))   (2)

where,

V dd is positive source voltage

V ss is negative source voltage

V Gate(0) is voltage across the power MOSFET 38 gate at the beginning of the switching period

M is mutual inductance, M=kL

k is inductor coupling factor

C eq is equivalent gate capacitance of the power MOSFET

Still referring to FIG. 3 and FIG. 4 , the next time interval begins when the power MOSFET 38 starts turning-on at t≈100 ns. At this time, drain voltage of the switching device Q 2 changes and becomes negative while current through Q 2 approaches zero. Since currents in the inductors L 2 and L 4 pass through Q 2 and can not disappear, the summary current I L2 +I L4 begins flowing through the diode D 1 thereby returning energy stored in the inductors L 2 and L 4 to the power source V dd . As the power MOSFET 38 turns-on, the gate drive current 10 begins recharging its reverse transfer (“Miller”) capacitance. A major part of this current comes from the source V dd through the inductors L 1 and L 3 . The recharging of the Miller capacitance also causes a dip in the power MOSFET 38 gate voltage waveform V gate 44 .

Still referring to FIG. 3 and FIG. 4 , the turn-off processes are similar to the ones that take place during the turn-on interval. At the beginning of the turn-off interval at t≈500 ns, the gate signal V gate Q 1 from VG 1 for the switching device Q 1 is switched-off, and it starts turning-off. The switching device Q 2 is not turned-on yet, so the turn-off interval begins with another “dead time” interval 43 when gate signals to both switching devices Q 1 and Q 2 are off. At t≈500 ns, the output of the gate drive circuit 10 (to the power MOSFET 38 gate) is positive, D 2 is off, and D 1 conducts current through inductors L 1 , L 2 , and Q 1 . As Q 1 turns-off, its drain voltage begins to rise and current in the inductor L 1 diverts to the inductor L 3 . At the same time, current in the inductor L 2 starts flowing through the power MOSFET 38 gate, the bias power source V dd (thereby regenerating energy), and the diode D 1 . Thus, at the end of the dead time interval at t≈550 ns, the power MOSFET 38 gate capacitance is partially discharged, switching devices Q 1 and Q 2 are off, currents in both diodes D 1 and D 2 are close to zero, and both coupled inductors L 1 , L 2 and L 3 , L 4 conduct current.

The dead time interval 43 ends at t≈550 ns when the gate signal V gate Q 2 from VG 2 starts turning-on the switching device Q 2 . As Q 2 turns-on, the voltage is applied to the coupled inductors L 1 , L 2 and L 3 , L 4 and the power MOSFET 38 gate is resonantly discharged by the sum of two currents. The source voltage V ss generates the first current that flows through Q 2 and L 2 in parallel with L 4 . The second current is the one that is stored in inductors L 1 and L 3 , and it flows through the integral body diode of the switching device Q 1 returning energy to the source V dd . At the end of the positive half wave of the resonant current I gate 46 the current in the integral body diode of the switching device Q 1 changes direction and starts flowing through the switching device's channel while it turns off. While Q 1 conducts current, the source V dd builds up current in inductors L 1 and L 3 . At the end of the power MOSFET 38 gate discharge time interval (t≈600 ns), the gate capacitance is almost fully discharged, Q 2 is on, Q 1 is off, and inductor currents continue increasing. Similarly to the turn-on, expressions (1) and (2) determine approximate values of the discharge current and its resonant frequency.

The next time interval begins when the power MOSFET 38 starts turning off at t≈600 ns. At this time, drain voltage of the switching device Q 1 turns negative and its current declines. Since currents in the inductors L 1 and L 3 flowing through Q 1 can not disappear, the summary current I L1 +I L3 flows through the diode D 2 thereby returning energy stored in the inductors L 1 and L 3 to the power source V ss . As the power MOSFET 38 turns off, the gate drive current begins recharging its reverse transfer (“Miller”) capacitance. A major part of this current comes from the voltage source V ss through the inductors L 2 and L 4 . The recharging of the Miller capacitance also causes a dip in MOSFET gate voltage waveform V gate 44 .

Referring to FIG. 2 , the energy regenerative gate drive circuit 10 has positive and negative voltage sources V dd , V ss . The negative source V ss improves noise immunity of the gate drive circuit 10 by developing negative voltage across the power MOSFET 38 gate during the off time. Both voltage sources take part in energy recovery wherein during turn-on of the MOSFET, energy is returned to the negative voltage source V ss , and during turn-off it is returned to the positive voltage source V dd . Therefore, the efficiency of the regenerative gate drive circuit 10 is improved.

During the dead time intervals 41 , 43 , gate capacitance of the power MOSFET 38 is partially discharged prior to on-to-off and off-to-on transitions thereby speeding up both switching transitions compared to the circuit disclosed in U.S. Pat. No. 5,264,736. The partial discharge represents MOSFET gate energy recovery through clamp diodes D 1 and D 2 and resonant inductors L 1 , L 2 and L 3 , L 4 .

Referring to FIG. 5 and FIG. 6 , FIG. 5 is a graph showing a measured power MOSFET 38 gate voltage relative to input gate signal VG 1 generated by prior art gate drive circuit of FIG. 1 , and FIG. 6 is a graph showing a measured power MOSFET 38 gate voltage relative to the input gate signal VG 1 , generated by the regenerative gate drive circuit 10 of FIG. 2 . The LLCC network 33 comprised of coupled inductors L 3 , L 4 and capacitors C 1 , C 2 speed-up charging and discharging power MOSFET 38 that is critically important to reducing switching losses. Compared with the gate drive circuit of FIG. 2 , the conventional gate driver disclosed in U.S. Pat. No. 5,264,736 has slower transitions. The LLCC network 33 also reduces the output impedance of the gate driver circuit 10 thereby improving its noise immunity, and it reduces the linear operation losses of the gate driver circuit 10 because gate drive switching devices Q 1 and Q 2 are decoupled from the power MOSFET 38 gate capacitance.

The prior art conventional gate driver of FIG. 1 has 33% higher power consumption than that of the gate driver circuit 10 of FIG. 2 . This was determined by averaging measurements of four prototypes of the gate driver circuit 10 at 750 kHz switching frequency.

Referring again to FIG. 2 the circuit components used to implement the preferred embodiment at the 1 MHz switching frequency are as follows: MOSFET Q 1 and MOSFET Q 2 may be embodied by part number IRFR 014 manufactured by International Rectifier of El Segundo, Calif.; Diodes D 1 and D 2 may be embodied by part no. MBR0540T1 manufactured by ON SEMICONDUCTOR of Phoenix, Ariz.; the completed inductors L 1 , L 2 and L 3 , L 4 may be embodied by part No. S33018 (self-inductance of each half-winding is 300 nH, K=0.92 manufactured by Vanguard Electronics, of Gardena, Calif.

One of ordinary skill in the art will recognize that if a bipolar transistor was used instead of MOSFETs, Q 1 and Q 2 , then an anti-parallel diode is required across each of Q 1 and Q 2 with its cathode connected to the drain.

This invention has been disclosed in terms of a certain embodiment. It will be apparent that many modifications can be made to the disclosed apparatus without departing from the invention. Therefore, it is the intent of the appended claims to cover all such variations and modifications as come within the true spirit and scope of this invention.