US20170019015A1 - Switching regulator with controllable dead time and quick start - Google Patents
Switching regulator with controllable dead time and quick start Download PDFInfo
- Publication number
- US20170019015A1 US20170019015A1 US14/798,903 US201514798903A US2017019015A1 US 20170019015 A1 US20170019015 A1 US 20170019015A1 US 201514798903 A US201514798903 A US 201514798903A US 2017019015 A1 US2017019015 A1 US 2017019015A1
- Authority
- US
- United States
- Prior art keywords
- delay
- dead time
- quick start
- side switch
- signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 claims description 56
- 238000012544 monitoring process Methods 0.000 claims description 33
- 230000004044 response Effects 0.000 claims description 32
- 230000000630 rising effect Effects 0.000 claims description 25
- 230000008859 change Effects 0.000 claims description 20
- 239000000872 buffer Substances 0.000 description 19
- 230000007704 transition Effects 0.000 description 15
- 238000010586 diagram Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 10
- 230000001934 delay Effects 0.000 description 9
- 230000008878 coupling Effects 0.000 description 8
- 238000010168 coupling process Methods 0.000 description 8
- 238000005859 coupling reaction Methods 0.000 description 8
- 238000010200 validation analysis Methods 0.000 description 6
- 238000013459 approach Methods 0.000 description 5
- 230000001276 controlling effect Effects 0.000 description 5
- 230000007423 decrease Effects 0.000 description 5
- 230000003111 delayed effect Effects 0.000 description 5
- 230000001105 regulatory effect Effects 0.000 description 5
- 239000003990 capacitor Substances 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000012546 transfer Methods 0.000 description 4
- 230000003247 decreasing effect Effects 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 230000001052 transient effect Effects 0.000 description 2
- XUIMIQQOPSSXEZ-UHFFFAOYSA-N Silicon Chemical compound [Si] XUIMIQQOPSSXEZ-UHFFFAOYSA-N 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 230000002939 deleterious effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000005669 field effect Effects 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 229910044991 metal oxide Inorganic materials 0.000 description 1
- 150000004706 metal oxides Chemical class 0.000 description 1
- 230000003071 parasitic effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 229910052710 silicon Inorganic materials 0.000 description 1
- 239000010703 silicon Substances 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02M—APPARATUS FOR CONVERSION BETWEEN AC AND AC, BETWEEN AC AND DC, OR BETWEEN DC AND DC, AND FOR USE WITH MAINS OR SIMILAR POWER SUPPLY SYSTEMS; CONVERSION OF DC OR AC INPUT POWER INTO SURGE OUTPUT POWER; CONTROL OR REGULATION THEREOF
- H02M1/00—Details of apparatus for conversion
- H02M1/36—Means for starting or stopping converters
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02M—APPARATUS FOR CONVERSION BETWEEN AC AND AC, BETWEEN AC AND DC, OR BETWEEN DC AND DC, AND FOR USE WITH MAINS OR SIMILAR POWER SUPPLY SYSTEMS; CONVERSION OF DC OR AC INPUT POWER INTO SURGE OUTPUT POWER; CONTROL OR REGULATION THEREOF
- H02M1/00—Details of apparatus for conversion
- H02M1/08—Circuits specially adapted for the generation of control voltages for semiconductor devices incorporated in static converters
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02M—APPARATUS FOR CONVERSION BETWEEN AC AND AC, BETWEEN AC AND DC, OR BETWEEN DC AND DC, AND FOR USE WITH MAINS OR SIMILAR POWER SUPPLY SYSTEMS; CONVERSION OF DC OR AC INPUT POWER INTO SURGE OUTPUT POWER; CONTROL OR REGULATION THEREOF
- H02M1/00—Details of apparatus for conversion
- H02M1/38—Means for preventing simultaneous conduction of switches
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02M—APPARATUS FOR CONVERSION BETWEEN AC AND AC, BETWEEN AC AND DC, OR BETWEEN DC AND DC, AND FOR USE WITH MAINS OR SIMILAR POWER SUPPLY SYSTEMS; CONVERSION OF DC OR AC INPUT POWER INTO SURGE OUTPUT POWER; CONTROL OR REGULATION THEREOF
- H02M3/00—Conversion of dc power input into dc power output
- H02M3/02—Conversion of dc power input into dc power output without intermediate conversion into ac
- H02M3/04—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters
- H02M3/10—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters using discharge tubes with control electrode or semiconductor devices with control electrode
- H02M3/145—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal
- H02M3/155—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal using semiconductor devices only
- H02M3/156—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal using semiconductor devices only with automatic control of output voltage or current, e.g. switching regulators
- H02M3/158—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal using semiconductor devices only with automatic control of output voltage or current, e.g. switching regulators including plural semiconductor devices as final control devices for a single load
- H02M3/1588—Conversion of dc power input into dc power output without intermediate conversion into ac by static converters using discharge tubes with control electrode or semiconductor devices with control electrode using devices of a triode or transistor type requiring continuous application of a control signal using semiconductor devices only with automatic control of output voltage or current, e.g. switching regulators including plural semiconductor devices as final control devices for a single load comprising at least one synchronous rectifier element
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
- Y02B70/00—Technologies for an efficient end-user side electric power management and consumption
- Y02B70/10—Technologies improving the efficiency by using switched-mode power supplies [SMPS], i.e. efficient power electronics conversion e.g. power factor correction or reduction of losses in power supplies or efficient standby modes
Definitions
- This disclosure relates generally to switching regulators and, more particularly, to switching regulator driver circuitry and techniques for dead time and quick start control.
- Regulators, or converters, including one or more switches, sometimes referred to as power switch(es), for transferring energy from an input, such as an AC or DC voltage or current source, to a regulated output are well known.
- the switch turns on and off to regulate the output.
- the switch operates in its active, or saturation region.
- Common switching regulator configurations include Buck, Boost, Buck-Boost, flyback, SEPIC, C ⁇ k, half bridge, and full bridge to name a few.
- various control methodologies for controlling conduction of the power switch(es) can be applied to switching regulators, including Pulse Width Modulation (PWM) and Pulse Frequency Modulation (PFM), and for each of these control methodologies, various feedback and feed forward techniques are possible including voltage mode control and current mode control.
- Switching regulators are often used to provide a regulated current and/or voltage to drive a load.
- Conventional Buck regulators generally contain a switch that conducts to transfer energy to an inductor during a first portion of each cycle and a diode (sometimes referred to as a freewheeling diode) that conducts during a second portion of each cycle to cause energy to be transferred from the inductor to the load.
- a diode sometimes referred to as a freewheeling diode
- a variation of this conventional Buck regulator is referred to as a synchronous Buck regulator in which the freewheeling diode is replaced with a second switch.
- this configuration requires additional control circuitry to drive both of the switches, use of the second switch can result in improved efficiency.
- FIG. 1 An example synchronous Buck regulator 10 is shown in FIG. 1 to include switches 12 , 14 , here in the form of n-channel metal oxide semiconductor field effect transistors (MOSFETs or simply FETs), coupled between an input voltage source VIN 13 and a reference such as ground 15 .
- a node SW 11 between the high side switch 12 and the low side switch 14 is coupled to an inductor 16 which stores energy for transfer to a load (not shown) coupled to an output terminal 18 at which a regulated output voltage VOUT is provided.
- An output capacitor 20 is also coupled to the output terminal 18 .
- the control circuit 48 includes an error amplifier 26 that is responsive to the VOUT voltage 18 and to a reference signal 24 to generate a COMP signal 66 across a series-coupled resistor 28 and capacitor 30 .
- the COMP signal 66 is coupled to an input of a comparator 32 that further receives a ramp signal from a summation element 36 . More particularly, the summation element 36 sums a first ramp signal 64 with a feedback signal 40 that is proportional to the current through the switches 12 , 14 .
- An output of the comparator 32 provides a reset input to a flip-flop 34 that is set by a clock signal 58 from an oscillator 38 .
- the output of the flip-flop 34 provides a feedback control signal HSON 52 that establishes on and off times of the high side switch 12 through the driver circuit 50 , here in the form of a buffer 46 , based on the output voltage VOUT 18 .
- the HSON signal 52 is additionally coupled to a delay element 45 and a buffer 44 to establish on and off times for the low side switch 14 , as shown.
- control circuit 48 and driver circuit 50 can be provided in an integrated circuit (IC) package and the remainder of the regulator circuitry 51 can be external to the IC package.
- IC integrated circuit
- the IC may be referred to generally as a driver IC.
- transitions between switching states be performed quickly with as little switching delay as possible since a slow slew rate and/or significant dead time (i.e., time when neither transistor is on, such as the time between the low side switch turning off and the high side switch turning on) can negatively impact regulator efficiency. It can be challenging to establish a switch slew rate and/or dead time that strikes an optimal balance between these competing requirements.
- Slew rate is a function of various factors, such as the switch gate impedance, the switch capacitance, and the load current.
- the slew rate can be “tuned” by using external resistors, such as resistors 70 , 72 in FIG. 1 , in series with the respective gate connection to establish well controlled gate impedance.
- resistors 70 , 72 in FIG. 1 in series with the respective gate connection to establish well controlled gate impedance.
- smaller gate impedance is necessary for larger FETs (i.e., FETs with larger switch capacitance that drive larger loads) and larger gate impedance is necessary for smaller FETs (i.e., FETs with smaller switch capacitance that drive smaller loads).
- the difficulty of implementing such switch slew rate optimization can be compounded by use of a driver IC to drive a variety of external switches for a variety of loads, since the switch capacitance and load current can vary.
- Dead time is affected by various parameters of the driven switches 12 , 14 , such as the threshold voltages, gate capacitance, and gate resistance.
- optimization of the dead time is challenging when such parameters are not well known or tightly controlled, such as when using a driver IC to drive a range of external FET switches 12 , 14 .
- this optimization can be even more challenging since these FET parameters are influenced by other factors, such as load current, input voltage, output voltage, and temperature variations.
- a driver circuit for driving a switching transistor of a switching regulator having a low side switch and a high side switch coupled to the low side switch at a switch node includes an adjustable delay element coupled between a control terminal of the high side switch and a control terminal of the low side switch and having a delay and a dead time calibration circuit configured to calibrate a dead time between the low side switch or the high side switch turning off and the other one of the high side switch or the low side switch turning on.
- the dead time calibration circuit includes a delay comparator configured to compare a present delay between the low side switch or the high side switch turning off and a voltage at the switch node to a past delay between the low side switch or the high side switch turning off and the voltage at the switch node and a controller responsive to the comparison by the delay comparator to adjust the delay of the adjustable delay element.
- the dead time calibration circuit is configured to calibrate the dead time between the low side switch turning off and the high side switch turning on and the delay comparator is configured to compare the present delay between the low side switch turning off and the voltage at the switch node rising to a past delay between the low side switch turning off and the voltage at the switch node rising.
- the dead time calibration circuit is configured to calibrate the dead time between the high side switch turning off and the low side switch turning on and the delay comparator is configured to compare the present delay between the high side switch turning off and the voltage at the switch node falling to a past delay between the high side switch turning off and the voltage at the switch node falling.
- the delay comparator may be configured to generate a first indicator when the present delay is longer than the past delay, a second indicator when the present delay is shorter than the past delay, and a third indicator when the present delay and the past delay are substantially the same.
- the controller may be configured to change the delay of the adjustable delay element in response to the first indicator and in response to the second indicator.
- the controller may be configured to hold the delay of the adjustable delay element constant in response to the third indicator.
- the delay comparator may be responsive to a low side control signal coupled to the control terminal of the low side switch or to a high side control signal coupled to the control terminal of the high side switch and to the voltage at the switch node.
- the controller may be further responsive to a fault signal to disable adjustment of the adjustable delay element.
- monitoring the delay comprises monitoring the delay between the low side switch turning off and the voltage at the switch node rising and in another embodiment, monitoring the delay comprises monitoring the delay between the high side switch turning off and the voltage at the switch node falling.
- Setting the dead time adjustment to an initial value may correspond to a maximum dead time target.
- Setting the dead time adjustment may include selecting a delay of a delay element coupled between a control terminal of the high side switch and a control terminal of the low side switch.
- Changing the dead time adjustment may include incrementing or decrementing the delay of the delay element.
- Monitoring the delay may include monitoring the voltage at the switch node and a low side control signal coupled to the low side switch or monitoring the voltage at the switch node and a high side control signal coupled to the high side switch.
- Monitoring the delay may include determining when the delay stops changing and an optimal dead time adjustment may be selected based on when the delay stops changing.
- the method may include validating the optimal dead time adjustment.
- validating the optimal dead time adjustment may include one or both of incrementing or decrementing the dead time adjustment and monitoring the delay between the low side switch turning off or the high side switch turning off and the voltage at the switch node.
- the method may include periodically rechecking the optimal dead time adjustment.
- driver circuit for driving a switching transistor of a switching regulator having a low side switch and a high side switch coupled to the low side switch at a switch node
- a quick start signal generator comprising an adjustable delay element to generate a quick start signal having a pulse to turn on the switching transistor for a quick start interval that terminates when the quick start pulse terminates, a quick start comparator configured to monitor the quick start signal, and a control circuit responsive to the comparison by the quick start comparator to adjust the delay of the delay element.
- the quick start comparator may be configured to determine if a dead time when neither the high side switch nor the low side switch is on in response to the quick start signal is greater than an upper limit related to a maximum dead time.
- the maximum dead time is a predetermined value.
- the quick start comparator may be configured to determine if the quick start pulse terminates after a lower limit related to the high side switch turning on and, in an embodiment, the quick start comparator may determine if the quick start pulse terminates after a voltage at the switch node rises.
- the quick start comparator may be configured to monitor the quick start signal to generate a first indicator if a dead time when neither the high side switch nor the low side switch is on in response to the quick start signal is greater than an upper limit related to a maximum dead time, a second indicator if the quick start pulse terminates after a lower limit related to the high side switch turning on, and a third indicator if both the dead time is less than the upper limit and the quick start pulse terminates before the lower limit.
- the control circuit may change the delay of the delay element in response to the first indicator and in response to the second indicator and/or may hold the delay of the delay element constant in response to the third indicator.
- the switching transistor is the high side switch and the driver circuit further includes a plurality of driver stages, each having a control input responsive to a respective driver control signal, wherein each of the driver control signals is a function of the quick start signal so that each of the plurality of driver stages is on during the quick start interval.
- a first one of the plurality of driver stages may be responsive only to the quick start signal and a second one of the plurality of driver stages may be responsive to a feedback signal received by the driver circuit and to the quick start signal so that the second one of the plurality of driver stages is still on after the quick start interval terminates.
- a method for generating a drive signal for controlling a switching transistor of a switching regulator having a low side switch and a high side switch coupled to the low side switch at a switch node including setting a pulse width of a quick start signal that controls the switching transistor during a quick start interval that terminates when the quick start pulse terminates, monitoring the quick start signal to determine if one or both of: (i) a dead time when neither the high side switch nor the low side switch is on in response to the quick start signal is greater than an upper limit related to a maximum dead time, or (ii) the quick start pulse terminates after a lower limit related to the high side switch turning on, and adjusting the pulse width of the quick start signal if one or both of: (i) the dead time is greater than the upper limit, or (ii) the quick start pulse terminates after the lower limit.
- the quick start signal is monitored to determine if both: (i) the dead time is greater than the upper limit and (ii) the quick start pulse terminates after the lower limit.
- the pulse width of the quick start signal may be held constant if the dead time is less than the upper limit and the quick start pulse terminates before the lower limit. Monitoring the quick start signal to determine if the dead time is less than the upper limit may include comparing the dead time to a predetermined maximum dead time value. Monitoring the quick start signal to determine if the quick start signal terminates after the lower limit may include comparing termination of the quick start pulse to a rising edge of the switch node voltage.
- the pulse width of the quick start signal may be set by setting a delay of an adjustable delay element to an initial value. Adjusting the pulse width of the quick start signal may include changing the delay of the delay element.
- FIG. 1 is a schematic of a conventional switching regulator
- FIG. 2 is a schematic of a switching regulator having a multi-stage driver circuit and a quick start controller
- FIG. 3 is a schematic of the multi-stage driver circuit of FIG. 2 ;
- FIG. 4 shows several illustrative waveforms associated with the regulator of FIG. 2 when the multi-stage driver drives switching transistors of different sizes
- FIG. 5 is a schematic of an alternative switching regulator having a multi-stage driver and a switch monitor circuit
- FIG. 6 is a schematic of the multi-stage driver of FIG. 5 ;
- FIG. 7 is a schematic of a switching regulator having a dead time calibrator
- FIG. 8 shows several illustrative waveforms associated with the regulator of FIG. 7 different dead times
- FIG. 9 is a schematic of an example delay compare circuit of FIG. 7 ;
- FIG. 10 is a flow diagram illustrating a method of calibrating the dead time with the dead time calibrator of FIG. 7 ;
- FIG. 11 is a schematic of an alternative switching regulator having a dead time calibrator and a quick start controller
- FIG. 12 shows several illustrative waveforms associated with the regulator of FIG. 11 for different operating conditions
- FIG. 13 is a schematic of an example quick start comparator of FIG. 11 ;
- FIG. 14 is a schematic of the quick start driver of FIG. 11 ;
- FIG. 15 is a flow diagram illustrating a method of calibrating the quick start interval with the quick start controller of FIG. 11 ;
- FIG. 16 is a schematic of another alternative switching regulator having a dead time calibrator and a quick start controller.
- a switching regulator 100 includes control circuit 48 , a driver circuit 80 , and regulator circuitry 90 .
- the driver circuit 80 differs from the driver circuit 50 ( FIG. 1 ) in that a multi-stage driver circuit 104 replaces the buffer 46 and the regulator additionally includes a quick start controller 106 that generates a quick start adjustment signal adjQstart 96 for coupling to the multi-stage driver 104 .
- the multi-stage driver circuit 104 provides a controlled slew rate to the high side switch 12 that is tailored to the size of the FEBT (i.e., to the switch capacitance).
- the quick start adjustment signal 96 establishes an initial driver interval (referred to herein alternatively as a quick start interval) during which the switch 12 is quickly charged to a point at or near its Miller plateau and also serves to initiate operation of the multi-stage driver circuit 104 .
- the regulator circuitry 90 differs from the converter circuitry 51 of FIG. 1 in that resistors 70 , 72 are eliminated, as is made possible by operation of the multi-stage driver circuit 104 .
- an example multi-stage driver circuit 104 includes a driver control signal generator 152 and driver stage circuitry 150 .
- the driver stage circuitry 150 includes a plurality of driver stages 102 a - 102 d , each having a control input responsive to a respective driver control signal 140 a - 140 d generated by the driver control signal generator 152 and an output coupled to the output of the other ones of the plurality of driver stages and to the control terminal 112 (i.e., the gate terminal) of the high side switch 12 ( FIG. 2 ).
- each driver stage 102 a - 102 d includes a driver transistor, such as in the form of the illustrated PMOS FETs, having a control input provided by its gate terminal and an output provided by its source terminal.
- the driver transistors 102 a - 102 d are coupled in parallel with their source terminals coupled together and their drain terminals coupled together, as shown.
- Each driver control signal 140 a - 140 d has an on time during which the respective driver transistor is on and an off time during which the respective driver transistor is off. At least one of the driver control signals 140 a - 140 d has an on time that is delayed with respect to an on time of another one of the driver control signals.
- the driver stages 102 a . . . 102 d are sequentially turned on to establish a controlled slew rate. As more driver stages are turned on, the total impedance of the parallel driver stages decreases in order to thereby allow the gate terminal 112 of the high side switch 12 to charge more quickly than otherwise possible.
- a slew time interval of a switch such as the high side switch 12 ( FIG. 2 ) can be described as commencing when the gate to source voltage of the switch reaches its Miller plateau and ending when the source to drain voltage reaches its final voltage level.
- the final voltage level of the source to drain voltage of switch 12 is substantially equal to the input voltage VIN 13 .
- the slew time interval coincides generally with the interval during which the switch node voltage SW 11 ( FIG. 2 ) rises.
- the Miller plateau is an effect that is often observed during transitions between a FETs off and on states that manifests itself as a flattening of the gate to source voltage during a portion of the transition. It is caused by the transistor's gate to drain parasitic capacitance (Cgd) pushing down on the gate voltage while the source and/or drain voltage slews, which causes the gate to source voltage to temporarily slow down or cease the increase (or decrease) in gate voltage (appearing as a plateau on the Vgs versus time plot).
- the Miller plateau voltage is a function of the transistor's threshold voltage, the application's load current, and the transistor's on resistance.
- At least two of the driver stages 102 a - 102 d are on during the slew time interval of the switch 12 .
- the multi-stage driver 104 provides decreasing impedance at the gate 112 of the switch 12 during the slew time interval in order to thereby speed up charging the gate to drain capacitance resulting in a change in the slew rate of the source to drain voltage.
- the driver control signal generator 152 includes at least one delay element, and here three delay elements 130 a - 130 c , to generate the driver control signals 140 a - 140 d , in response to the HSON signal 52 and a quick start signal 108 .
- the quick start adjustment signal 96 is coupled to control an adjustable delay element 92 .
- the output of delay element 92 and the HSON signal 52 are coupled to AND gate 94 that generates the quick start signal 108 , as shown.
- a logic gate 132 receives the HSON signal 52 and the quick start signal 108 and provides an output signal to start the first delay element 130 a on the falling edge of the quick start signal.
- the output of the first delay element 130 a is coupled to an input of the second delay element 130 b and the output of the second delay element 130 b is coupled to an input of the third delay element 130 c , as shown.
- the first delay element 130 a provides a 12 ns delay
- the second delay element 130 b provides a 7 ns delay
- the third delay element 130 c provides a 7 ns delay.
- Logic gate 128 a receives the HSON signal 52 and logic gates 128 b - 128 d receive the outputs of the three delay elements 130 a - 130 c , respectively, and each gate 128 a - 128 d additionally receives the quick start signal 108 .
- This configuration results in the first driver stage ( 102 a ) being on when the HSON signal 52 is active.
- the driver control signal generator 152 is implemented on a “low side” of the regulator 100 in the sense that its signal levels are logic signal levels; whereas, the driver stage circuitry 150 is on a “high side” of the regulator, with the SW node signal level ranging from VIN 13 ( FIG. 2 ) to ground.
- the BOOT signal 114 may be provided by a bootstrap capacitor (not shown) connected between BOOT and the switch node SW 11 in order to maintain the required gate drive voltage. The bootstrap capacitor may be charged when the SW node 11 approaches ground, such as with a diode connected to a voltage source.
- Level shifters 126 a - 126 d are coupled between respective outputs of OR gates 128 a - 128 d and the driver stage circuitry 150 in order to translate the logic level signals associated with the OR gates 128 a - 128 d to high side signal levels for coupling to the driver stage circuitry 150 .
- the driver control signal generator 152 is implemented on the low side
- this signal generating circuitry could alternatively be implemented on the high side, thereby reducing the required number of level shifters 126 a - 126 d .
- the decision to implement the signal generator 152 on the low side versus the high side impacts silicon area, ESD protection, and other factors depending on the process. For example implementing delay cells 130 a - 130 c on the high side can result in a larger physical circuit for each delay cell, but to implement them on the low side requires more level shifters, which also requires more physical area.
- the driver stage circuitry 150 includes pre-driver buffer stages 124 a - 124 d and 122 a - 122 d .
- the buffer stages are sized to achieve a predetermined gate drive level for the driver stages 102 a - 102 d and each buffer stage may have the same or different drive capability. It will be appreciated that additional or fewer pre-driver buffer stages may be provided.
- the buffers 122 a - 122 d and 124 a - 124 d are provided in the form of inverters.
- One of the level shifted signals here a signal from level shifter 126 a that corresponds to the delayed signal 140 a that controls the first driver stage 102 a , is coupled to a buffer inverter 120 for further coupling to a buffer inverter 118 and to a gate terminal of a NMOS FET 110 .
- the NMOS FET 110 has a drain terminal coupled to the gate terminal 112 of the high side switch 12 ( FIG. 2 ) and a source terminal coupled to the switch node SW 11 ( FIG. 2 ). In operation, the NMOS transistor 110 is off when any of the PMOS driver transistors 102 a - 102 d is on.
- FIG. 4 shows an example HSON signal 52 and quick start signal 108 . Also shown in FIG. 4 are the driver control signals 140 a - 140 c , the gate current Igate 160 of the high side switch 12 ( FIG. 2 ), the gate to source voltage, HSgate 164 of the switch 12 , and the switch node voltage SW 11 .
- Each of the illustrated signals is shown in connection with driving three different high side switches 12 , each having a different size and therefore different input capacitance.
- a first portion 204 of the waveforms illustrates the respective signals when the multi-stage driver circuit 104 is coupled to a relatively small high side FET 12 with a relatively small input capacitance
- a second portion 206 of the waveforms illustrates the signals when the multi-stage driver circuit 104 is coupled to a medium sized FET 12
- a third portion 208 of the waveforms illustrates the signals when the multi-stage driver circuit 104 is coupled to a relatively large FET 12 .
- the quick start controller 106 ( FIG. 2 ) are described below in connection with quick start control and calibration circuit 510 of FIG. 11 . Suffice it to say here that under certain operating conditions (e.g., when driving a transistor 12 of a certain size), the quick start signal 108 (via OR gates 128 a - 128 d ) turns on all of the gate drive stages 102 a - 102 d initially, at the rising edge of the HSON signal 52 , for a quick start interval that ends when the gate to source voltage, HSgate 164 is at or near (but not above) the Miller plateau as is labeled 252 for the medium sized FET 206 and 254 for the large FET 208 .
- the quick start signal 108 via OR gates 128 a - 128 d ) turns on all of the gate drive stages 102 a - 102 d initially, at the rising edge of the HSON signal 52 , for a quick start interval that ends when the gate to source voltage, HSgate 164 is at or
- the quick start signal 108 is provided in the form of a pulse that commences in response to a transition of the HSON signal 52 and ends when the switch node voltage SW 11 starts to rise. In this way, the end of the quick start pulse or interval can be considered to provide an indicator of the start of the slew time interval and/or the transistor being at or near its Miller plateau. Use of the quick start signal 108 to quickly charge the transistor to a point at or near its Miller plateau can reduce the total duration of the transition of the gate to source voltage Vgs of the transistors.
- driver control signal 140 a stays high rather than going low to turn off driver stage 102 a .
- the quick start pulse 108 turns on all of the drive stages 102 a - 102 d with its rising edge and turns off all but the first drive stage 102 a with its falling edge.
- the second through fourth driver stages are turned on by respective driver control signals 140 b - 140 d according to their respective delays, as shown.
- the first portion 204 of the waveforms of FIG. 4 corresponds to driving a high side FET 12 that is so small that it does not receive a quick start pulse 108 .
- the first driver control signal 140 a transitions to a high state to turn on the first driver stage 102 a .
- a first delay time later, here 12 ns later as established by delay element 130 a the second driver control signal 140 b goes high to turn on the second driver stage 102 b .
- the third driver control signal 140 c goes high to turn on the third driver stage 102 c and finally, after a third delay time of 7 ns as established by delay element 130 c , the fourth driver control signal 140 d goes high to turn on the fourth driver stage 102 d.
- driver stages 102 a - 102 d may be turned on during the slew time interval depending on the capacitance of the driver FET. For example, in the case of the smallest FET, only the first driver stage 102 a is required to charge the SW node 11 to the VIN voltage level (i.e., only one driver stage 102 a is on during the slew time interval).
- the first and second driver stages 102 a , 102 b are required in order to charge the SW node 11 to the VIN voltage level (i.e., two driver stages 102 a , 102 b are on during the slew time interval).
- a dotted line waveform labeled 256 illustrates what the switch node voltage SW 11 would look like if only a single driver stage (e.g., 102 a ) were on during the slew time interval.
- a dotted line waveform labeled 260 illustrates what the switch node voltage SW 11 would look like if only the first and second driver stages 102 a , 102 b were on during the slew time interval and a dotted line waveform labeled 258 illustrates what the switch node voltage SW 11 would look like if only the first driver stage 102 a were on during the slew time interval.
- the delay elements 130 a - 130 c may provide the same delays or different delays.
- Various considerations may be used to select the delays provided by the delay elements in order to ensure that a predetermined slew time interval is achieved for all possible sizes of driven FETs.
- the delays can be selected to achieve the predetermined slew time interval for the smallest driven FET with only a first driver transistor 102 a on and to require some additional number of driver transistors to be on to achieve the same predetermined slew time interval for the largest driven FET.
- the first delay element 130 a establishes a longer delay (i.e., 12 ns) than the second and third delay elements 130 b , 130 c in order to provide some time for the first stage to drive the gate to source voltage to the Miller plateau, whether the quick start signal 108 is present or not.
- the quick start interval (when the quick start signal 108 is active) is not present during the Miller plateau.
- each stage is sequentially turned on such that a subsequent stage is turned on while the stage(s) that were previously on remain on in order to thereby achieve a total driver impedance equal to the parallel combination of the impedances of all of the active driver stages. It will be appreciated that other sequencing schemes are also possible. Further, while the example multi-stage driver 104 includes four driver stages 102 a . . . 102 d , other numbers of stages are possible.
- the impedance of the driver transistors 102 a - 102 d may be the same as each other or may be different in order to achieve a desired scaling of the total parallel impedance as driver stages are turned on.
- the impedance of driver stage 102 a is 30 ⁇
- the impedance of driver stage 102 b is 30 ⁇
- the impedance of driver stage 102 c is 15 ⁇
- the impedance of driver stage 102 d is 5 ⁇ .
- the total driver impedance is 30 ⁇
- the total driver impedance is 15 ⁇
- the total driver impedance is 7.5 ⁇
- the total driver impedance is 3 ⁇ .
- multiple driver stages 102 a - 102 d are only used for the pull up devices (i.e., the PMOS FETs) of the driver circuit 104 .
- the multi-stage driver approach can be applied to the pull down gate drive (i.e., NMOS FET 110 ) or to the low side switch (i.e., FET 14 ), which can be required by different switcher topologies, such as an inverted buck LED driver topology where the inductor current flows in the opposing direction of the illustrated embodiment.
- the multi-stage driver approach may also be desirable on the low side of an H-Bridge buck-boost regulator constructed with two switch nodes SW, with a high side multi-stage driver on the input switch node SW and a low side multi-stage driver on the output switch node SW, as will be described in connection with FIG. 16 .
- the pull up devices are focused on because controlling the slew rate of the SW rising edge can be more important than the falling edge. Since, during this transition, current is being sourced from the input supply VIN, which has a relatively large current loop, the slew rate of the rising edge of the switch voltage SW 11 can have more of an impact on EMI/EMC performance.
- the quick start signal 108 can also be used as a mechanism for initiating operation of the multi-stage driver 104 by operation of gate 132 ( FIG. 3 ). It will be appreciated that this functionality can be achieved with alternative circuits and techniques, such as an adaptive quick start period, a pre-programmable quick start period, or other methods.
- the multi-stage driver can be implemented without a quick start interval.
- One such alternative regulator 170 that does not include a quick start interval to more quickly charge the FET 12 to its Miller plateau is shown in FIG. 5 .
- Regulator 170 differs from the regulator 100 ( FIG. 2 ) in that the quick start controller 106 is replaced with a switch node monitor 114 that generates a swRising signal 116 and the multi-stage driver 104 is replaced with an alternative multi-stage driver circuit 180 that responds to the swRising signal 116 rather than to a quick start adjustment signal.
- the falling edge of quick start signal 108 is analogous to the rising edge of the swRising signal 116 .
- the falling edge of quick start signal 108 occurs before the start of the Miller plateau and the rising edge of the swRising signal 116 occurs after the Miller plateau.
- the swRising signal 116 is used to initiate operation of the multi-stage driver 104 . More particularly, the swRising signal 116 provides an indicator of the start of the slew time interval.
- switch node monitor 114 monitors the voltage at the switch node SW 11 to determine when the switch 12 has reached its Miller plateau and may include a comparator to compare the switch node voltage SW 11 to a reference voltage selected so that a transition of the swRising signal 116 (e.g., a transition to a logic high level) coincides generally to the switch 12 reaching its Miller plateau as can be determined by monitoring when the switch node voltage starts to rise.
- the swRising signal 116 is at a first logic level when the switch node voltage SW 11 is greater than the reference voltage and is at a second logic level when the switch node voltage SW 11 is less than the reference voltage.
- the switch node monitor comparator may include hysteresis.
- the reference voltage may be user programmable or otherwise programmable or adjustable. In order to better reject variations in the rise time of the switch node voltage SW 11 , it may be desirable to provide the reference voltage at a relatively low level, such as on the order of 1 volt for a 12 volt supply, or more generally, on the order of approximately 10% of the supply voltage VIN.
- the delay between the SW node voltage 11 rising edge and the swRising signal 116 is also desirable for the delay between the SW node voltage 11 rising edge and the swRising signal 116 to be as small as possible.
- the swRising signal could also be generated by monitoring the HSgate signal with a threshold configured below the Miller plateau; however, the Miller plateau is often an unknown voltage.
- the multi-stage driver circuit 180 includes a driver control signal generator 190 and driver stage circuitry 150 .
- the driver control signal generator 190 includes an AND gate 144 that receives the HSON signal 52 and the swRising signal 116 , as shown. Thus, the output of gate 144 transitions to a high level upon a rising edge of the swRising signal 116 and transitions to a low level once the HSON signal 52 transitions to a low level to indicate that the Miller plateau has been reached.
- the driver control signal generator 190 includes three delay elements, here elements 146 a - 146 c , each providing a respective delay.
- delay element 146 a establishes a delay of 5 ns
- delay element 146 b establishes a delay of 7 ns
- delay element 146 c establishes a delay of 7 ns.
- a first control signal 140 a turns on and off the first driver stage 102 a according to the timing of the HSON signal 52
- a second control signal 140 b turns on the second driver stage 102 b 5 ns after the swRising signal 116 rising edge (since the delay element 146 a is triggered by the swRising signal 116 going high)
- a third control signal 140 c turns on the third driver stage 102 c 7 ns after the second driver stage 102 b is turned on
- a fourth control signal 140 d turns on the fourth driver stage 102 d 7 ns after the third driver stage 102 c is turned on.
- the swRising signal 116 initiates operation of the multi-stage driver 180 in the same fashion as the quick start signal 108 ( FIG. 3 ) such that the delayed driver control signals commence at a controlled time following a transition of the swRising signal 116 that indicates the beginning of the slew time interval, but here without the initial quick start charging implemented by the quick start signal 108 ( FIG. 3 ).
- the delays established by delay elements 146 a - 146 c can be readily modified.
- a switching regulator 200 implementing a dead time calibration includes control circuit 48 and regulator circuitry 90 .
- the regulator 200 further includes a driver circuit 210 that differs from the driver circuit 80 ( FIG. 2 ) in that the fixed delay element 45 is replaced with an adjustable delay element 212 .
- a dead time control and calibration circuit, or simply control circuit, or calibrator 240 includes circuitry and implements techniques for optimizing the dead time between turning off the low side switch 14 and turning on the high side switch 12 by adjusting the delay established by the delay element 212 .
- the dead time calibrator 240 sets a dead time adjustment to an initial value and changes the dead time adjustment while monitoring a delay between the switch node voltage SW 11 rising and the low side switch 14 turning off to determine when to stop changing the dead time adjustment based on when the delay stops changing, as will be described. It will be appreciated that while the dead time calibration circuit is described in connection with certain embodiments ( FIGS. 7 and 11 for example) to optimize the dead time between turning off the low side switch 14 and turning on the high side switch, in other embodiments ( FIG.
- dead time calibration optimizes the dead time between turning off the high side switch and turning on the low side switch in which case the dead time calibrator sets a dead time adjustment to an initial value and changes the dead time adjustment while monitoring a delay between the switch node voltage falling and the high side switch turning off to determined when to stop changing the dead time adjustment based on when the delay stops changing.
- the switch monitor 114 compares the switch node voltage SW 11 to a reference voltage to generate the swRising signal 116 .
- the swRising signal 116 is at a first logic level when the switch node voltage SW 11 is greater than a reference voltage and is at a second logic level when the switch node voltage is less than the reference voltage.
- the reference voltage is relatively low, such as on the order of 1 volt for a 12 volt supply voltage VIN 13 . In this way, the swRising signal 116 provides an indication of the switch node voltage 11 rising.
- a delay comparator 214 receives the swRising signal 116 , an LSgate signal 56 coupled to the gate terminal of the low side switch 14 , and a sampleDead signal 230 and compares the swRising signal 116 to the LSgate signal 56 to determine a present delay between the voltages 116 , 56 and also to determine whether the present delay is greater than, less than, or equal to a previously determined delay (i.e., a past delay), as will be explained below in connection with the example delay comparator 214 of FIG. 9 .
- Outputs of the delay comparator 214 include a deadLonger signal 226 that, when high for example, indicates that the present delay is longer than the past delay and a deadShorter signal 228 that, when low for example, indicates that the present delay is shorter than the past delay. Additionally, delay comparator 214 drives both the deadLonger signal 226 and the deadShorter signal 228 low, for example, indicating that the present delay is substantially equal to the past delay.
- a dead time calibration state machine 244 responds to the deadLonger signal 226 and to the deadShorter signal 228 and implements a calibration routine whereby, the dead time adjustment is either decremented or incremented or unchanged. To this end, the state machine 244 generates an incDead signal 220 to increment the dead time adjustment and a decDead signal 222 to decrement the dead time adjustment via a dead register 216 .
- the delay of the delay element 212 is controlled by an address 218 (i.e., by a dead time adjustment 218 ) from the dead register 216 and, in an embodiment, can have a width on the order of 5 bits and a LSB weight of 1-3 ns for example.
- the dead register 216 generates and stores the address 218 with an up/down counter controlled by the incDead signal 220 and the decDead signal 222 .
- a resetDead signal 224 resets the register 216 to set the adjustable delay element 212 to an initial, minimum delay value, resulting in the most positive (i.e., maximum) dead time for the system.
- the dead time calibrator 240 initially sets the dead time adjustment to an initial value (here, to a minimum adjustment value) corresponding to an initial, maximum dead time.
- the calibrator 240 slowly attempts to decrement the dead time by incrementing the dead time adjustment by a small step, such as on the order of between 1 and 3 nanoseconds, following which the dead time is monitored to determine if a change in the delay between the LSgate signal 56 and the swRising signal 116 (i.e., a change in the dead time) has occurred. If a change in the dead time occurs, then the dead time adjustment is again incremented and the process is repeated.
- the dead time calibrator 240 operates on the principle that the optimal dead time occurs slightly before when the relative delay between the switch node voltage SW 11 and the low side switch control signal LSgate 56 no longer changes in response to a change (e.g., an increment) of the dead time adjustment. It will be appreciated that in embodiments (e.g., FIG. 16 ) in which the dead time calibration controls the dead time between the high side switch turning off and the low side switch turning on, the optimal dead time occurs slightly before when the relative delay between the switch node voltage and the high side switch control signal no longer changes in response to a change of the dead time adjustment. The reason that no further change in the dead time is indicative of being at or near the transition from a positive dead time to a negative dead time is because this condition occurs when the dead time is negative. In many applications, it is desirable to have a slight positive dead time.
- example signal waveforms associated with the regulator 200 of FIG. 7 include the switch node voltage SW 11 , the LSgate signal 56 , the high side switch drain to source current Ids 314 , and the high side switch gate to source voltage, or control signal HSgate 316 .
- Each of the illustrated signals is shown in connection with various dead times.
- a first portion 304 of the waveforms illustrates the respective signals when the regulator 200 is operating with several different dead times (with the left most waveforms corresponding to the most negative dead time), a second portion 306 of the waveforms illustrates the signals when the regulator 200 is operating with a first, positive dead time, a third portion 308 of the waveforms illustrates the signals when the regulator 200 is operating with a second, optimal dead time, and a fourth portion 310 of the waveforms illustrates the signals when the regulator 200 is operating with a third, negative dead time.
- the peak drain current 314 decreases with each step because the body diode of the low side switch 14 has less time for charge to be built up.
- the conduction of the low side switch 14 body diode results from the switch node voltage SW 11 going negative.
- the peak of drain current 314 significantly increases due to the high side switch 12 and the low side switch 14 simultaneously conducting (shoot through current).
- the switch node slew rate increases dramatically. This increase in slew rate is due to excess gate to source voltage 316 on the high side FET 12 when the low side switch 14 and the high side switch 12 conduct simultaneously.
- the dead time calibration routine selects as the optimal dead time a dead time that is one or two steps from the point in time (i.e., the dead time adjustment 218 ) where relative delay between the low side switch turning off and the switch node voltage rising no longer changes.
- an example delay comparator 214 is shown to receive the LSgate signal 56 and the swRising signal 116 and to provide the deadLonger signal 226 and the deadShorter signal 228 .
- the delay comparator 214 is additionally responsive to a sampleDead signal 230 from the dead time calibration state machine 244 ( FIG. 7 ).
- the delay comparator 214 measures and stores the present delay between the LSgate signal 56 and the swRising signal 116 when the sampleDead signal 230 is high and otherwise determines the relative delay (i.e., compares the present delay to a stored past delay). Suffice it to say here that the deadLonger signal 226 indicates when the present delay is longer than the past delay and the deadShorter signal 228 indicates when the present delay is shorter than the past delay.
- the LSgate signal 56 is coupled to an inverter 705 and to a delay line 700 including a plurality of delay cells 704 a - 704 y configured to provide a plurality of delay signals 728 to unit cells 702 a - 702 y .
- the adjustable delay element 212 and the delay comparator 214 are designed using similar delay cells so that shifts due to process, temperature, or bias conditions will tend to cancel.
- the delay comparator 214 has a resolution on the order of the least significant bit (LSB) weight of the adjustable delay element 212 ( FIG. 7 ).
- each delay cell 704 a - 704 y may have a delay of about one-half the LSB delay step of the adjustable delay element 212 ( FIG. 7 ). With this arrangement, the delay comparator 214 is able to reject small deviations in the delay comparison that are not attributable to the LSB delay step of the adjustable delay element 212 .
- the delay cells 704 a - 704 y can establish the same or different delays as each other. Using different magnitudes of delays can allow the delay comparator 214 to measure a wider range of delays than otherwise possible, since the total sum of the delay established by the delay line 700 corresponds to the maximum delay that the delay comparator 214 can measure.
- a first plurality of the delay cells 704 a - 704 y can provide a first predetermined delay and a second plurality of the delay cells can provide a second predetermined delay that is shorter than the first predetermined delay.
- Each unit cell 702 a - 702 y includes a first register 714 , as may be implemented with a latch, to store the present delay and provide a present delay signal 726 and a second register 716 , as also may be implemented with a latch, to store the past delay and provide a past delay signal 728 .
- the present delay register 714 is strobed every period that the switch node 11 is active (i.e. during the slew time interval) by a strbDead signal 730 .
- the strbDead signal 730 is generated in response to the LSgate signal 56 , thereby latching the register on the rising edge of swRising signal 116 , and the swRising signal 116 and in one example, corresponds to the dead time when neither transistor is on.
- the duration of the strbDead signal 730 defines the interval during which the present delay between the LSgate signal 56 and the swRising signal 116 is measured and stored.
- the sampleDead signal 230 ( FIG. 7 ) is buffered by a buffer 708 to provide a strbReg signal 732 that enables the past register 716 to transfer the present delay from the present delay register 714 to the past delay register 716 . This transfer may occur when the strbDead signal 730 is low (i.e., when the present delay register 714 is disabled).
- Logic gates 718 and 720 are coupled to receive the present delay signal 726 and the past delay signal 728 , as shown.
- the output signals 734 from gates 718 of the plurality of unit cells 702 a - 702 y are coupled through further gates 722 , 710 to generate the deadShorter signal 228 for coupling to the dead time calibration state machine 244 ( FIG. 7 ).
- the output signals 736 from the gates 720 from the plurality of unit cells 702 a - 702 y are coupled through further gates 724 , 712 to generate the deadLonger signal 226 for coupling to the dead time calibration state machine 244 ( FIG. 7 ).
- the deadShorter signal 228 or for the deadLonger signal 226 to be at a logic high level at least two consecutive unit cells 702 a - 702 y must have a mismatch between the past delay signal 728 and the present delay signal 726 . More particularly, when the present delay signal 726 is low and the past delay signal 728 is high, then the output signal 734 will be at a logic high, and if two consecutive outputs 734 are high, logic gate 722 will pass a logic high signal through OR gate 710 forcing the deadShorter signal 228 to be at a logic high level.
- the described digital approach may reject process variations and mismatches better than some other techniques, such as analog techniques.
- a flow diagram illustrates a technique 400 implemented by the dead time calibrator 240 of FIG. 7 .
- the rectangular elements (typified by element 402 ) are herein denoted “processing blocks” and the diamond-shaped elements (typified by element 404 ) are herein denoted “decision blocks” and either or both may represent computer software instructions or groups of instructions.
- processing blocks the rectangular elements
- diamond-shaped elements are herein denoted “decision blocks” and either or both may represent computer software instructions or groups of instructions.
- FIG. 10 (and other flow diagrams herein) represent exemplary embodiments of designs disclosed herein and variations in such embodiments, which generally follow the processes outlined, are considered to be within the scope of the concepts, systems and techniques described and claimed herein. Some or all of the blocks may represent operations performed by functionally equivalent circuits.
- the dead time is initially set to a maximum value in block 401 by the resetDead signal 224 .
- the dead register 216 may be set to a value corresponding to a minimum delay by delay element 212 in order to achieve a maximum dead time target, since for a delay coupled to the low side switch 14 , an attempt to decrement the dead time is equivalent to increasing the delay provided by delay element 212 .
- a dead time decrementing phase 440 is entered during which it is attempted to decrement the dead time at block 402 , following which the relative delay between the switch node voltage SW 11 and the LSgate signal 56 is determined at decision block 404 .
- the relative delay (difference between the present delay and past delay) is captured by transferring the present delay from the present delay register 714 to the past delay register 716 ( FIG. 9 ) in response to strbReg signal 732 ( FIG. 9 ) as indicated by the “sampleDead” statement in blocks 402 , 408 , 414 , 420 , and 421 .
- the dead time is again attempted to be decremented at block 402 .
- the dead time is again attempted to be decremented at block 402 following a wait period at block 406 .
- the present delay being shorter than the past delay indicates that the dead time is positive.
- the wait period block 406 may be on the order of sixty switching cycles (set by the bandwidth of the regulation loop) to ensure that the calibration routine does not respond to transients.
- the delay compare block 404 and wait block 406 alternatively could be implemented by averaging the delay comparison over several cycles and proceeding to attempt to decrement the dead time if a majority of the cycles yields a shorter relative delay.
- a dead time tuning and validation phase 442 is entered. More particularly, validation is performed at decision block 418 to validate that an optimal dead time has been achieved. No change in the relative delay following the dead time decrement block 402 indicates that the dead time has crossed from being a positive dead time to a negative dead time. Once the dead time is negative, further attempts to decrement the dead time will not cause a change in the result of the delay comparator 214 ( FIG. 7 ), since the low side switch 14 solely determines when the switch node SW 11 is released.
- any further attempts to decrement the dead time at this point will increase the shoot through current and the switch node slew rate, due to overcharging the high side gate voltage HSgate 316 as shown in FIG. 8 .
- the delay element 212 ( FIG. 7 ) is immediately decremented (i.e., the incDead signal 220 is incremented in block 414 without a wait state) in order to impact the efficiency as little as possible.
- the dead time corresponding to the point at or very near to when the dead time crosses from being positive to negative can be considered an optimal dead time.
- the dead time may be incremented at block 414 as may be desirable to back off slightly from the dead time setting at which a positive dead time just becomes a negative dead time in order to minimize the efficiency loss associated with a negative dead time.
- the wait state 416 causes the validation 418 to be performed over many cycles, such as on the order of 60 cycles, to ensure that transients do not affect the determination.
- the optimal dead time has not been found. More particularly, if the present delay is determined to be longer than the past delay ( 430 ), then the calibrator returns to block 402 to again attempt to decrement the dead time. Alternatively, if the present delay is determined to be shorter than the past delay ( 428 ), then the calibrator attempts to increase the dead time (and additionally captures the relative delay) at block 408 and again determines at decision block 411 whether the relative delay has changed. If it is determined that the present delay is longer than the past delay, it is again attempted to decrement the dead time at block 402 .
- the dead time is again incremented at block 408 .
- This additional comparison block 411 ensures that a positive dead time is achieved so that the decrementing phase 440 can find the optimal dead time.
- validation block 418 If it is determined in validation block 418 that the relative delay has not changed, then the validation has passed ( 432 ). In this case, a further increment of the dead time may be made in block 420 in order to further increase the efficiency, following which a dead time calibration termination phase 446 may be entered. More particularly, the relative delay is captured in block 421 by the sampleDead signal 230 going high so that subsequent changes in the dead time can be detected.
- the system idles at block 412 and continuously checks the relative delay at decision block 410 in order to determine whether a change in the system operating conditions results in the optimal dead time changing.
- a determination at decision block 410 that the present delay is shorter than the past delay ( 434 ) can indicate that the threshold voltage of the switches 12 , 14 has changed resulting in a more negative dead time. Accordingly, the dead time is attempted to be incremented at block 408 until the present dead time becomes longer than the past dead time. If it is determined at decision block 410 that the present delay is longer than the past delay, then the dead time has become more positive due to changing operating conditions and the decrementing phase 440 recommences, as shown.
- the dead time calibration method 400 presumes that the switch node 11 is switching and that the regulator operating conditions have not dramatically changed.
- a resetCal signal 234 and a holdCal signal 232 are provided to the state machine ( FIG. 7 ), as may be from an external control unit or processor.
- the resetCal signal 234 is forced high during system events that significantly change the operating point of the switch node SW 11 , such as start up of the regulator or fault events.
- the state machine 244 is forced into and held at an initial state 401 which also forces the resetDead signal 224 high to reset the dead register 216 to the initial, most positive dead time setting.
- the holdCal signal 232 is used.
- the holdCal signal 232 is high, the contents of the dead register 216 remain unchanged and the state machine 244 is held at a wait or comparison state (e.g., states 404 , 406 , 416 , 418 , 411 , 421 , 412 , or 410 ).
- a wait or comparison state e.g., states 404 , 406 , 416 , 418 , 411 , 421 , 412 , or 410 .
- the dead time calibrator 240 is described as adjusting only the delay between the high side control signal HSON 52 and the low side control signal LSgate 56 , in some embodiments, it may be desirable to additionally or alternatively control a delay associated with driving the high side switch 12 .
- the above circuitry and techniques can additionally include a further delay element between the HSON signal 52 and the gate terminal of the high side switch 12 with which the high side control signal HSgate 316 can be delayed relative to the low side control signal LSgate 56 under certain operating conditions, such as when the delay element 212 sets the maximum initial dead time (i.e., corresponding to a minimum delay) or during a predetermined range of dead times including the maximum initial dead time.
- a switching regulator 500 that implements a quick start calibration in addition to dead time calibration includes control circuit 48 , regulator circuitry 90 , a driver circuit 504 , dead time control and calibration circuit, or calibrator 240 , and a quick start control and calibration circuit, or simply control circuit, or calibrator.
- the driver circuit 504 includes a quick start driver 512 that is responsive to the HSON signal 52 and generates a high side control signal HSgate 508 for coupling to the gate terminal of the high side switch 12 .
- the driver 512 is further responsive to a quick start adjustment adjQstart signal 522 from the quick start control circuit 510 and provides a quickStart signal 520 to the quick start control circuit 510 , as shown.
- the quickStart signal 520 may be the same as or similar to the quick start signal 108 ( FIG. 3 ).
- the quick start signal 520 establishes an initial driver interval (referred to herein alternatively as the quick start interval) during which the switch 12 is quickly charged to at or near its Miller plateau, without overshoot.
- the quickStart signal 520 can be provided in the form of a pulse that commences in response to a transition of the low side control signal LSgate 56 and ends when the switch node voltage SW 11 starts to rise. In this way, the end of the quick start signal, or pulse can be considered to provide an indicator of the start of the slew time interval. It will be appreciated that an optimal quick start interval will vary based on the threshold voltage and the gate capacitance of the high side switch 12 . Because of this variation, the calibrator 510 includes circuitry and implements techniques to tailor the quick start interval to the switch parameters.
- Quick start calibrator 510 includes a state machine 514 , a quick start comparator 516 , and a Qstart register 518 .
- the quick start comparator 516 monitors the quickStart signal 520 and the state machine 514 adjusts the quickStart signal via the adjQstart signal 522 under certain conditions.
- the quick start comparator 516 is configured to determine if a dead time (i.e., a time when neither the high side switch 12 nor the low side switch 14 is on) resulting from the quickStart signal 520 is greater than an upper limit related to a maximum dead time target.
- the quick start comparator 516 may additionally or alternatively monitor the quickStart signal 520 to determine if the quick start pulse terminates after a lower limit related to the high side switch 12 turning on (i.e., when the switch node voltage 11 begins to rise).
- An example quick start comparator 516 is shown in FIG. 13 . Suffice it to say here that the comparator 516 generates a deadLongError signal 530 that provides an indication of whether the dead time exceeds the upper limit and a deadShortError signal 532 that provides an indication of whether the quickStart signal pulse terminates after the lower limit.
- both the deadLongError signal 530 and the deadShortError signal 532 are low if the dead time is less than the upper limit and the quick start pulse terminates before the lower limit, the deadLongError signal 530 is high and the deadShortError signal 532 is low if the dead time is greater than the upper limit, and the deadLongError signal 530 is low and the deadShortError signal 532 is high if the quickStart pulse terminates after lower limit.
- the state machine 514 performs a quick start calibration routine to generate an incQstart signal 524 , a decQstart signal 526 , and a resetQstart signal 528 , all of which are coupled to the Qstart register 518 .
- the Qstart register 518 generates the adjQstart signal 522 with an up/down counter in response to the incQstart signal 524 , the decQstart signal 526 , and/or the resetQstart signal 528 to control a delay in the driver 512 .
- the address bus 522 has a width of 5 bits and results in a LSB weight of 0.3 ns to 3 ns, for example.
- An example driver 512 is shown in FIG. 14 and described below.
- example signal waveforms associated with the regulator 500 of FIG. 11 include the switch node voltage 11 , the high side switch drain to source current Ids 314 , the high side control signal HSgate 508 , the low side control signal LSgate 56 , and the quickStart signal 520 .
- Each of the illustrated signals is shown in connection with various operating conditions.
- a first portion 574 of the waveforms illustrates the respective signals when the regulator 500 is operating with several different quick start intervals decreasing from left to right
- a second portion 576 of the waveforms illustrates the signals when the regulator 500 is operating with a first quick start interval
- a third portion 578 of the waveforms illustrates the signals when the regulator 500 is operating with a second, optimal quick start interval
- a fourth portion 580 of the waveforms illustrates the signals when the regulator 500 is operating with a third quick start interval.
- Waveform potion 576 illustrates an operating point when the quick start interval is considered too short, as can result in a dead time that is greater than a maximum dead time target (i.e., greater than an upper limit) necessary to achieve a relatively short switch on time target. This condition is flagged by the deadLongError signal 530 ( FIG. 11 ).
- Waveform portion 578 illustrates an optimal quick start interval as is apparent by the quick start interval stopping just before the Miller plateau 592 is reached, thereby allowing a higher impedance gate drive to control the switch node as shown in FIG. 14 .
- the last portion 580 of the waveforms illustrates that when the quick start interval is too long, as can result when the quick start interval terminates after the switch node voltage SW 11 begins to rise (i.e., when the quick start pulse terminates after the lower limit), the gate to source voltage HSgate 508 can overshoot, resulting in a fast switch node rise time. This condition is flagged by the deadShortError signal 532 ( FIG. 11 ).
- an example quick start comparator 516 includes a first delay element 554 that is responsive to an inverted version of the ISgate signal 56 provided by an inverter 550 to generate an input signal for a latch 558 .
- the delay element 554 sets an upper limit for the regulator dead time and may be selected based on the minimum on time of the SW node 11 , such as on the order of 10 ns for a 75 ns minimum on time. More particularly, the delay established by delay element 554 may be selected to provide a relatively small permissible margin beyond a predetermined maximum dead time target.
- a second delay element 560 responsive to the quickStart signal 520 sets a lower limit for the quick start interval and provides an input signal to a latch 570 .
- the delay established by delay element 560 may be selected to provide a relatively small permissible margin beyond high side switch turning on (i.e., when the switch node voltage SW 11 begins to rise), but during which the quick start pulse still may be active. Therefore, delay element 560 accounts for the inherent delay associated with translating the control signals on the low side to the output (i.e., to the gate terminal 112 of the high side switch) on the high side.
- Latches 558 and 570 are enabled (i.e., strobed) from a time when the low side switch 14 turns off to a time when the switch node voltage SW 11 begins to rise thereby latching the state on the swRising signal 116 rising edge, as is achieved in response to the strbDead signal 568 from a logic gate 564 .
- the output of latch 558 is the deadLongError signal 530 that is indicative of the dead time being greater than the upper limit.
- the output of latch 570 is the deadShortError signal 532 that is indicative of the quickStart signal 520 being asserted outside of the strobe period (i.e., the quick start interval ending after the lower limit when the switch node voltage begins to rise).
- an example quick start driver circuit 512 includes a driver control signal generator 640 and driver stage circuitry 644 .
- the driver stage circuitry 644 includes a plurality of driver stages 660 a - 660 b , each having a control input responsive to a respective driver control signal 662 a - 662 b generated by the driver control signal generator 640 and an output coupled to the output of the other ones of the plurality of driver stages and to the control terminal (i.e., the gate terminal) of the high side switch 12 ( FIG. 11 ).
- each driver stage 660 a - 660 b includes a driver transistor, such as in the form of the illustrated PMOS FETs, having a control input provided by its gate terminal and an output provided by its drain terminal.
- the driver transistors 660 a , 660 b are coupled in parallel with their source terminals coupled together and their drain terminals coupled together, as shown.
- Each driver control signal 662 a - 662 b has an on time during which the respective driver transistor is on and an off time during which the respective driver transistor is off. At least one of the driver control signals 662 a - 662 b has an on time that is controlled by the quickStart signal 520 and at least another one of the driver control signals has an on time controlled by a combination of the quickStart signal 520 and the HSON signal 52 .
- both of the driver stages 660 , 660 b are on; whereas, after the quick start interval, only one of the driver stages 660 a , 660 b is on.
- both of the driver stages 660 a , 660 b are on during the quick start interval and thereafter, only one of the driver stages, 660 b is on during the slew time interval of the switch 12 (as the switch node voltage SW 11 rises).
- the driver control signal generator 640 includes an adjustable delay element 650 to generate the quickStart signal 520 in response to the HSON signal 52 and the adjQstart signal 522 .
- a logic gate 652 receives the HSON signal 52 and an output signal from the adjustable delay element 650 and generates the quickStart signal 520 .
- Logic gate 654 a receives the HSON signal 52 and logic gates 654 a - 654 b receive the quickStart signal 520 , as shown.
- Level shifters 664 a - 664 b are coupled between respective outputs of OR gates 654 a - 654 b and the driver stage circuitry 644 in order to translate the logic level signals associated with the OR gates 654 a - 654 b to high side signal levels for coupling to the driver stage circuitry 644 .
- the driver control signal generator 640 could alternatively be implemented on the high side, thereby reducing the number of required level shifters 664 a - 664 b.
- the driver stage circuitry 644 includes pre-driver buffer stages 678 a - 678 b , 680 a - 680 b .
- the buffer stages are sized to achieve a predetermined gate drive level for the driver stages 660 a - 660 b and each buffer stage may have the same or different drive capability. It will be appreciated that additional or fewer pre-driver buffer stages may be provided.
- the buffers 678 a - 678 b and 680 a - 680 b are provided in the form of inverters.
- One of the level shifted signals here a signal from level shifter 664 a that corresponds to the delayed signal 662 a that controls the first driver stage 660 a , is coupled to a buffer inverter 672 for further coupling to a buffer inverter 674 and to a gate terminal of an NMOS FET 670 .
- the NMOS FET 670 has a drain terminal coupled to the gate terminal of the high side switch 12 ( FIG. 11 ) and a source terminal coupled to the switch node SW 11 ( FIG. 11 ). In operation, the NMOS transistor 670 is off when any of the PMOS driver transistors 660 a - 660 b is on.
- a flow diagram illustrates a technique 600 implemented by the quick start calibrator 510 of FIG. 11 .
- the Qstart register 518 is reset in response to the resetQstart signal 528 . It is then determined at block 602 , such as by comparator 516 ( FIG. 11 ), whether a dead time when neither the high side switch 12 nor the low side switch 14 is on that occurs in response to the quickStart signal 520 is greater than an upper limit related to a maximum dead time target and whether the quick start signal pulse terminates after a lower limit related to the high side switch turning on.
- the quick start signal is considered to be within an acceptable range and the calibration routine proceeds to a wait state 604 in order to ensure that the result from decision block 602 was not due to a transient.
- Illustrative wait intervals implemented by wait block 604 may be on the order of 60 PWM cycles.
- the function of comparison by the quick start comparator 516 in block 602 and the wait state 604 could alternatively be implemented by averaging the comparison over several cycles and proceeding forward with the comparison result that has the highest average.
- the state machine 514 ( FIG. 11 ) then rechecks the quickStart signal pulse width again to determine if it is within the comparison window (i.e., if the dead time is less than the upper limit and the quick start pulse terminates before the lower limit) with decision block 610 . If it is determined at decision block 610 that the quick start signal 520 within the comparison window ( 632 ), then the quick start calibration is terminated at block 614 . If it is determined at decision block 610 that the dead time is greater than the upper limit (as may be indicated by the deadLongError signal 530 ), then the quick start interval is incremented at block 606 and processing continues as described above.
- the quick start pulse terminates after the lower limit (as may be indicated by signal deadShortError signal 532 ), then the quick start interval is decremented at block 608 and processing continues as described above.
- the Qstart register 518 is incremented by the incQstart signal 524 at block 606 .
- the quick start signal 522 is incremented, it is determined at block 612 whether the Qstart register 516 is at the upper limit. If the Qstart register 516 is not at the upper limit (0), then a wait block 611 is entered, following which the process repeats beginning with block 602 . In this way, the quickStart signal pulse width is repeatedly incremented for as long as the Qstart register 516 is not at the upper limit. If alternatively, it is determined at block 612 that the Qstart register 516 is at the upper limit (1), then the quick start calibration is deemed completed at block 614 because no further quick start can be applied.
- the quick start signal is again monitored at block 620 . If in block 620 , the quick start pulse is determined to terminate after the lower limit ( 634 ), then the process returns to block 608 to decrement the pulse width. If alternatively it is determined at block 620 that the dead time is greater than the upper limit or that the both the dead time is less than the upper limit and the quick start pulse terminates before the lower limit ( 636 ), then the calibration routine is completed at block 614 .
- the Qstart register 518 is decremented by the decQstart signal 526 at block 608 .
- the quick start calibration includes decrementing the quickStart signal pulse width in response to an active deadShortError signal ( 624 ) since this signal indicates that the end of the quick start interval is too close to the Miller plateau and could result in the switch node voltage SW 11 rising too fast. If the deadShortErrror persists even after the Qstart register 518 is at its lower limit, as done after decision block 602 , then the calibration is “permanently” completed in block 618 .
- a QcalDone signal 536 and a recalQstart signal 538 may be coupled between the quick start calibration state machine 514 and the dead time calibration state machine 244 (in embodiments including both calibrations).
- the quick start calibration routine 600 must complete before the dead time calibration routine 400 ( FIG. 10 ) starts. This is because the quick start calibration will significantly alter the dead time possibly making it impractical to run the dead time calibration routine at the same time as the quick start calibration routine.
- the calibration must be performed with the most positive dead time, which is when the dead register 216 ( FIG. 11 ) is held in reset.
- the QcalDone signal 536 may be used to initiate the start of the dead time calibration routine. At any of the blocks where the quick start calibration is deemed completed (blocks 614 , 618 ), the QcalDone signal 536 is asserted. In addition, if the dead time calibration routine cannot complete (e.g., because the dead time is at its limit in state 402 ), the quick start calibration routine 600 is signaled to restarted with the recalQstart signal 538 .
- the quick start calibration (like the dead time calibration described above) requires that the switch node 11 be switching and that the regulator operating conditions not dramatically change.
- the resetCal signal 234 and the holdCal signal 232 are provided to the quick start state machine 514 ( FIG. 11 ), as may be from an external control unit or processor.
- the resetCal signal 234 is forced high during system events that significantly change the operating point of the switch node SW 11 , such as start up of the regulator or fault events.
- the state machine 514 is forced into and held at an initial state 601 which also forces the resetDead signal 224 high to reset the Qstart register 518 .
- the holdCal signal 232 is used.
- the holdCal signal 232 is high, the contents of the Qstart register 518 remain unchanged and the state machine 514 is held at a wait or comparison state (e.g., states 602 , 610 , 620 , 604 , and 611 ). Once the holdCal signal 232 goes low, the process 600 resumes.
- the quick start signal 522 is described as controlling the duration of the quick start period, this signal may alternatively or additionally be used to control the drive impendence.
- the quick start period would be a fixed duration, 10 ns for example, and the adjQstart signal 552 would control the drive strength used during the fixed quick start period.
- the adjustable drive strength may be implemented with multiple drivers using a variety of on/off combinations to generate a variety of drive strengths.
- an H-Bridge buck boost regulator 700 implementing quick start calibration and a dead time calibration includes control circuit 48 , regulator circuitry 702 , an input driver circuit 504 a , an output driver circuit 504 b , a dead time and quick start calibration circuit 704 a to control the input driver circuit 504 a and a dead time and quick start calibration circuit 704 b to control the output driver circuit 504 b .
- the H-bridge regulator circuitry 702 includes an input high side switch 12 a , an input low side switch 14 a , an output high side switch 12 b , and an output low side switch 14 b , coupled to inductor 16 as shown to generate a regulated output voltage Vout 18 .
- the dead time and quick start calibration circuit 704 a can be the same as or similar to a combination of dead time calibrator 240 and quick start calibrator 510 ( FIG. 11 ).
- the output driver 504 b differs from driver 504 a in that a delay element 212 b (that may be the same as or similar to delay element 212 of FIG. 11 ) is coupled to the gate of high side output switch 12 b through buffer 44 b and the quick start driver circuit 512 b (that may be the same as or similar to quick start driver circuit 512 of FIG. 11 ) is coupled to the gate of the low side output switch 14 b .
- the dead time and quick start calibration circuit 704 b can thus, be the same as or similar to a combination of dead time calibrator 240 and quick start calibrator 510 of FIG. 11 , but is configured to control delay element 212 b coupled to the high side output switch 12 b and to control quick start driver 512 b coupled to the low side output switch 14 b , as shown.
- the dead time calibrator 240 FIG. 11
- the calibrator 704 b is configured to calibrate the dead time between the high side switch 12 b turning off and the low side switch 14 b turning on.
- the calibration circuit 704 a may be configured to monitor the delay between the low side switch 14 a turning off and the voltage at the switch node SWa rising in order to determine when to stop changing the dead time adjustment; whereas the calibration circuit 704 b may be configured to monitor the delay between the high side switch 12 b turning off and the voltage at the switch node SWb falling in order to determine when to stop changing the dead time adjustment.
- the quick start calibration implemented by circuits 704 a , 704 b may be combined in the sense that only one such circuit (circuit 704 a for example) may generate the adjQstart signal (signal 522 a for example) and that same adjQstart signal 522 a may be coupled to both the quick start driver 512 a and also to the quick start driver 512 b .
- the optimized quick start signal 522 a for example may also be used to drive the high side switch of a secondary high side switch (not shown) in the same system regulating a second load.
- the dead time calibration implemented by circuits 704 a , 704 b can be combined in the sense that only one such circuit (circuit 504 a for example) may generate the adjDead signal (signal 218 a for example) and that same adjDead signal 218 a may be coupled to both the delay element 212 a and also to the delay element 212 b.
- the proposed circuitry and techniques can be applied to any linear or switching regulator topology including but not limited to Buck, Boost, Buck-Boost, SEPIC, C ⁇ k, half-bridge, full bridge, and linear regulators utilizing with any type of control loop including current mode control, voltage mode control, constant on time control, constant off time control, or any other analog and/or digital control scheme.
- the proposed circuitry and techniques can be applied to regulators that regulate current, voltage, power, or other parameters.
- the circuitry and techniques described herein can be implemented using hardware, software, and/or firmware in a digital and/or analog fashion.
- certain terms used herein, such as controller, processor, control circuit, state machine can be implemented in any suitable fashion and are not intended to require any particular implementation methodology.
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Dc-Dc Converters (AREA)
Abstract
Description
- Not Applicable.
- Not Applicable.
- This disclosure relates generally to switching regulators and, more particularly, to switching regulator driver circuitry and techniques for dead time and quick start control.
- Regulators, or converters, including one or more switches, sometimes referred to as power switch(es), for transferring energy from an input, such as an AC or DC voltage or current source, to a regulated output are well known. In some regulators, sometimes referred to as switching regulators, the switch turns on and off to regulate the output. In other regulators, sometimes referred to as linear regulators, the switch operates in its active, or saturation region.
- Common switching regulator configurations include Buck, Boost, Buck-Boost, flyback, SEPIC, Cúk, half bridge, and full bridge to name a few. As is also well known, various control methodologies for controlling conduction of the power switch(es) can be applied to switching regulators, including Pulse Width Modulation (PWM) and Pulse Frequency Modulation (PFM), and for each of these control methodologies, various feedback and feed forward techniques are possible including voltage mode control and current mode control. Switching regulators are often used to provide a regulated current and/or voltage to drive a load.
- Conventional Buck regulators generally contain a switch that conducts to transfer energy to an inductor during a first portion of each cycle and a diode (sometimes referred to as a freewheeling diode) that conducts during a second portion of each cycle to cause energy to be transferred from the inductor to the load. A variation of this conventional Buck regulator is referred to as a synchronous Buck regulator in which the freewheeling diode is replaced with a second switch. Although this configuration requires additional control circuitry to drive both of the switches, use of the second switch can result in improved efficiency.
- An example synchronous Buck
regulator 10 is shown inFIG. 1 to includeswitches voltage source VIN 13 and a reference such asground 15. Anode SW 11 between thehigh side switch 12 and thelow side switch 14 is coupled to aninductor 16 which stores energy for transfer to a load (not shown) coupled to anoutput terminal 18 at which a regulated output voltage VOUT is provided. Anoutput capacitor 20 is also coupled to theoutput terminal 18. - Conduction of
switching transistors control circuit 48 anddriver circuit 50. In the illustratedregulator 10, thecontrol circuit 48 includes anerror amplifier 26 that is responsive to theVOUT voltage 18 and to areference signal 24 to generate aCOMP signal 66 across a series-coupledresistor 28 andcapacitor 30. TheCOMP signal 66 is coupled to an input of acomparator 32 that further receives a ramp signal from asummation element 36. More particularly, thesummation element 36 sums afirst ramp signal 64 with afeedback signal 40 that is proportional to the current through theswitches comparator 32 provides a reset input to a flip-flop 34 that is set by aclock signal 58 from anoscillator 38. The output of the flip-flop 34 provides a feedbackcontrol signal HSON 52 that establishes on and off times of thehigh side switch 12 through thedriver circuit 50, here in the form of abuffer 46, based on theoutput voltage VOUT 18. TheHSON signal 52 is additionally coupled to adelay element 45 and abuffer 44 to establish on and off times for thelow side switch 14, as shown. - In some embodiments, the
control circuit 48 anddriver circuit 50 can be provided in an integrated circuit (IC) package and the remainder of theregulator circuitry 51 can be external to the IC package. In this type of arrangement, the IC may be referred to generally as a driver IC. - It is desirable to operate switching regulators in a manner that enhances the electromagnetic compatibility (EMC) performance of the regulator. For example, in applications where a driver IC is used to drive one or more external transistors, it is generally desirable that the switching activity of the external transistor(s) cause as little electromagnetic interference (EMI) as possible to surrounding circuitry. It is known that fast slew rate (i.e., the rate of change of the switch drain to source, Vds, voltage per unit time) can contribute to EMI/EMC problems. However, it is also desirable that transitions between switching states be performed quickly with as little switching delay as possible since a slow slew rate and/or significant dead time (i.e., time when neither transistor is on, such as the time between the low side switch turning off and the high side switch turning on) can negatively impact regulator efficiency. It can be challenging to establish a switch slew rate and/or dead time that strikes an optimal balance between these competing requirements.
- Slew rate is a function of various factors, such as the switch gate impedance, the switch capacitance, and the load current. In applications in which the switch capacitance and load current are well defined to within a relatively narrow range, the slew rate can be “tuned” by using external resistors, such as
resistors 70, 72 inFIG. 1 , in series with the respective gate connection to establish well controlled gate impedance. In general, in order to achieve substantially the same slew rate, smaller gate impedance is necessary for larger FETs (i.e., FETs with larger switch capacitance that drive larger loads) and larger gate impedance is necessary for smaller FETs (i.e., FETs with smaller switch capacitance that drive smaller loads). However, the difficulty of implementing such switch slew rate optimization can be compounded by use of a driver IC to drive a variety of external switches for a variety of loads, since the switch capacitance and load current can vary. - It is also desirable to operate switching regulators in a manner that optimizes the dead time. If the dead time is too long, the body diode of the
low side switch 14 will conduct, which decreases the regulator efficiency due to switching and conduction losses of the diode and the reverse recovery time associated with turning off the diode. On the other hand, a dead time that is too short can result in both thehigh side switch 12 and thelow side switch 14 being on at the same time, which can cause undesirable shoot through currents that can adversely impact EMI performance and efficiency. - Dead time is affected by various parameters of the driven
switches external FET switches - A driver circuit for driving a switching transistor of a switching regulator having a low side switch and a high side switch coupled to the low side switch at a switch node includes an adjustable delay element coupled between a control terminal of the high side switch and a control terminal of the low side switch and having a delay and a dead time calibration circuit configured to calibrate a dead time between the low side switch or the high side switch turning off and the other one of the high side switch or the low side switch turning on. The dead time calibration circuit includes a delay comparator configured to compare a present delay between the low side switch or the high side switch turning off and a voltage at the switch node to a past delay between the low side switch or the high side switch turning off and the voltage at the switch node and a controller responsive to the comparison by the delay comparator to adjust the delay of the adjustable delay element. With this arrangement, dead time optimization over a range of switch and system parameters is achieved and maintained. In an embodiment, the dead time calibration circuit is configured to calibrate the dead time between the low side switch turning off and the high side switch turning on and the delay comparator is configured to compare the present delay between the low side switch turning off and the voltage at the switch node rising to a past delay between the low side switch turning off and the voltage at the switch node rising. In another embodiment, the dead time calibration circuit is configured to calibrate the dead time between the high side switch turning off and the low side switch turning on and the delay comparator is configured to compare the present delay between the high side switch turning off and the voltage at the switch node falling to a past delay between the high side switch turning off and the voltage at the switch node falling.
- Features may include one or more of the following. The delay comparator may be configured to generate a first indicator when the present delay is longer than the past delay, a second indicator when the present delay is shorter than the past delay, and a third indicator when the present delay and the past delay are substantially the same. The controller may be configured to change the delay of the adjustable delay element in response to the first indicator and in response to the second indicator. The controller may be configured to hold the delay of the adjustable delay element constant in response to the third indicator. The delay comparator may be responsive to a low side control signal coupled to the control terminal of the low side switch or to a high side control signal coupled to the control terminal of the high side switch and to the voltage at the switch node. The controller may be further responsive to a fault signal to disable adjustment of the adjustable delay element.
- Also described is a method for calibrating a dead time of a switching regulator comprising a high side switch and a low side switch coupled to the high side switch at a switch node including setting a dead time adjustment to an initial value, changing the dead time adjustment, and monitoring a delay between the low side switch turning off or the high side switch turning off and a voltage at the switch node to determine when to stop changing the dead time adjustment. In an embodiment, monitoring the delay comprises monitoring the delay between the low side switch turning off and the voltage at the switch node rising and in another embodiment, monitoring the delay comprises monitoring the delay between the high side switch turning off and the voltage at the switch node falling.
- Features may include one or more of the following. Setting the dead time adjustment to an initial value may correspond to a maximum dead time target. Setting the dead time adjustment may include selecting a delay of a delay element coupled between a control terminal of the high side switch and a control terminal of the low side switch. Changing the dead time adjustment may include incrementing or decrementing the delay of the delay element. Monitoring the delay may include monitoring the voltage at the switch node and a low side control signal coupled to the low side switch or monitoring the voltage at the switch node and a high side control signal coupled to the high side switch. Monitoring the delay may include determining when the delay stops changing and an optimal dead time adjustment may be selected based on when the delay stops changing.
- The method may include validating the optimal dead time adjustment. In an embodiment, validating the optimal dead time adjustment may include one or both of incrementing or decrementing the dead time adjustment and monitoring the delay between the low side switch turning off or the high side switch turning off and the voltage at the switch node. The method may include periodically rechecking the optimal dead time adjustment.
- Also described is driver circuit for driving a switching transistor of a switching regulator having a low side switch and a high side switch coupled to the low side switch at a switch node including a quick start signal generator comprising an adjustable delay element to generate a quick start signal having a pulse to turn on the switching transistor for a quick start interval that terminates when the quick start pulse terminates, a quick start comparator configured to monitor the quick start signal, and a control circuit responsive to the comparison by the quick start comparator to adjust the delay of the delay element. With this arrangement, a quick start feature is implemented in a controlled fashion to achieve an optimal balance between a relatively fast turn on of the switching transistor without the deleterious effects of a fast switch node slew rate.
- Features may include one or more of the following. The quick start comparator may be configured to determine if a dead time when neither the high side switch nor the low side switch is on in response to the quick start signal is greater than an upper limit related to a maximum dead time. In an embodiment, the maximum dead time is a predetermined value. The quick start comparator may be configured to determine if the quick start pulse terminates after a lower limit related to the high side switch turning on and, in an embodiment, the quick start comparator may determine if the quick start pulse terminates after a voltage at the switch node rises.
- The quick start comparator may be configured to monitor the quick start signal to generate a first indicator if a dead time when neither the high side switch nor the low side switch is on in response to the quick start signal is greater than an upper limit related to a maximum dead time, a second indicator if the quick start pulse terminates after a lower limit related to the high side switch turning on, and a third indicator if both the dead time is less than the upper limit and the quick start pulse terminates before the lower limit. The control circuit may change the delay of the delay element in response to the first indicator and in response to the second indicator and/or may hold the delay of the delay element constant in response to the third indicator.
- In an embodiment, the switching transistor is the high side switch and the driver circuit further includes a plurality of driver stages, each having a control input responsive to a respective driver control signal, wherein each of the driver control signals is a function of the quick start signal so that each of the plurality of driver stages is on during the quick start interval. A first one of the plurality of driver stages may be responsive only to the quick start signal and a second one of the plurality of driver stages may be responsive to a feedback signal received by the driver circuit and to the quick start signal so that the second one of the plurality of driver stages is still on after the quick start interval terminates.
- Also described is a method for generating a drive signal for controlling a switching transistor of a switching regulator having a low side switch and a high side switch coupled to the low side switch at a switch node including setting a pulse width of a quick start signal that controls the switching transistor during a quick start interval that terminates when the quick start pulse terminates, monitoring the quick start signal to determine if one or both of: (i) a dead time when neither the high side switch nor the low side switch is on in response to the quick start signal is greater than an upper limit related to a maximum dead time, or (ii) the quick start pulse terminates after a lower limit related to the high side switch turning on, and adjusting the pulse width of the quick start signal if one or both of: (i) the dead time is greater than the upper limit, or (ii) the quick start pulse terminates after the lower limit. In an embodiment, the quick start signal is monitored to determine if both: (i) the dead time is greater than the upper limit and (ii) the quick start pulse terminates after the lower limit.
- Features may include one or more of the following. The pulse width of the quick start signal may be held constant if the dead time is less than the upper limit and the quick start pulse terminates before the lower limit. Monitoring the quick start signal to determine if the dead time is less than the upper limit may include comparing the dead time to a predetermined maximum dead time value. Monitoring the quick start signal to determine if the quick start signal terminates after the lower limit may include comparing termination of the quick start pulse to a rising edge of the switch node voltage. The pulse width of the quick start signal may be set by setting a delay of an adjustable delay element to an initial value. Adjusting the pulse width of the quick start signal may include changing the delay of the delay element.
- The foregoing features of the disclosure, as well as the disclosure itself may be more fully understood from the following detailed description of the drawings. The drawings aid in explaining and understanding the disclosed technology. Since it is often impractical or impossible to illustrate and describe every possible embodiment, the provided figures depict one or more exemplary embodiments. Accordingly, the figures are not intended to limit the scope of the invention. Like numbers in the figures denote like elements.
-
FIG. 1 is a schematic of a conventional switching regulator; -
FIG. 2 is a schematic of a switching regulator having a multi-stage driver circuit and a quick start controller; -
FIG. 3 is a schematic of the multi-stage driver circuit ofFIG. 2 ; -
FIG. 4 shows several illustrative waveforms associated with the regulator ofFIG. 2 when the multi-stage driver drives switching transistors of different sizes; -
FIG. 5 is a schematic of an alternative switching regulator having a multi-stage driver and a switch monitor circuit; -
FIG. 6 is a schematic of the multi-stage driver ofFIG. 5 ; -
FIG. 7 is a schematic of a switching regulator having a dead time calibrator; -
FIG. 8 shows several illustrative waveforms associated with the regulator ofFIG. 7 different dead times; -
FIG. 9 is a schematic of an example delay compare circuit ofFIG. 7 ; -
FIG. 10 is a flow diagram illustrating a method of calibrating the dead time with the dead time calibrator ofFIG. 7 ; -
FIG. 11 is a schematic of an alternative switching regulator having a dead time calibrator and a quick start controller; -
FIG. 12 shows several illustrative waveforms associated with the regulator ofFIG. 11 for different operating conditions; -
FIG. 13 is a schematic of an example quick start comparator ofFIG. 11 ; and -
FIG. 14 is a schematic of the quick start driver ofFIG. 11 ; -
FIG. 15 is a flow diagram illustrating a method of calibrating the quick start interval with the quick start controller ofFIG. 11 ; -
FIG. 16 is a schematic of another alternative switching regulator having a dead time calibrator and a quick start controller. - Referring to
FIG. 2 , aswitching regulator 100 includescontrol circuit 48, adriver circuit 80, andregulator circuitry 90. Thedriver circuit 80 differs from the driver circuit 50 (FIG. 1 ) in that amulti-stage driver circuit 104 replaces thebuffer 46 and the regulator additionally includes aquick start controller 106 that generates a quick startadjustment signal adjQstart 96 for coupling to themulti-stage driver 104. Themulti-stage driver circuit 104 provides a controlled slew rate to thehigh side switch 12 that is tailored to the size of the FEBT (i.e., to the switch capacitance). The quickstart adjustment signal 96 establishes an initial driver interval (referred to herein alternatively as a quick start interval) during which theswitch 12 is quickly charged to a point at or near its Miller plateau and also serves to initiate operation of themulti-stage driver circuit 104. Theregulator circuitry 90 differs from theconverter circuitry 51 ofFIG. 1 in thatresistors 70, 72 are eliminated, as is made possible by operation of themulti-stage driver circuit 104. - Referring also to
FIG. 3 , an examplemulti-stage driver circuit 104 includes a drivercontrol signal generator 152 anddriver stage circuitry 150. Thedriver stage circuitry 150 includes a plurality of driver stages 102 a-102 d, each having a control input responsive to a respective driver control signal 140 a-140 d generated by the drivercontrol signal generator 152 and an output coupled to the output of the other ones of the plurality of driver stages and to the control terminal 112 (i.e., the gate terminal) of the high side switch 12 (FIG. 2 ). In the illustrated embodiment, each driver stage 102 a-102 d includes a driver transistor, such as in the form of the illustrated PMOS FETs, having a control input provided by its gate terminal and an output provided by its source terminal. The driver transistors 102 a-102 d are coupled in parallel with their source terminals coupled together and their drain terminals coupled together, as shown. - Each driver control signal 140 a-140 d has an on time during which the respective driver transistor is on and an off time during which the respective driver transistor is off. At least one of the driver control signals 140 a-140 d has an on time that is delayed with respect to an on time of another one of the driver control signals. With this arrangement, the driver stages 102 a . . . 102 d are sequentially turned on to establish a controlled slew rate. As more driver stages are turned on, the total impedance of the parallel driver stages decreases in order to thereby allow the
gate terminal 112 of thehigh side switch 12 to charge more quickly than otherwise possible. - A slew time interval of a switch, such as the high side switch 12 (
FIG. 2 ), can be described as commencing when the gate to source voltage of the switch reaches its Miller plateau and ending when the source to drain voltage reaches its final voltage level. In theillustrated regulator 100, the final voltage level of the source to drain voltage ofswitch 12 is substantially equal to theinput voltage VIN 13. The slew time interval coincides generally with the interval during which the switch node voltage SW 11 (FIG. 2 ) rises. - The Miller plateau is an effect that is often observed during transitions between a FETs off and on states that manifests itself as a flattening of the gate to source voltage during a portion of the transition. It is caused by the transistor's gate to drain parasitic capacitance (Cgd) pushing down on the gate voltage while the source and/or drain voltage slews, which causes the gate to source voltage to temporarily slow down or cease the increase (or decrease) in gate voltage (appearing as a plateau on the Vgs versus time plot). The Miller plateau voltage is a function of the transistor's threshold voltage, the application's load current, and the transistor's on resistance.
- In an embodiment, at least two of the driver stages 102 a-102 d are on during the slew time interval of the
switch 12. With this arrangement, themulti-stage driver 104 provides decreasing impedance at thegate 112 of theswitch 12 during the slew time interval in order to thereby speed up charging the gate to drain capacitance resulting in a change in the slew rate of the source to drain voltage. - The driver
control signal generator 152 includes at least one delay element, and here three delay elements 130 a-130 c, to generate the driver control signals 140 a-140 d, in response to theHSON signal 52 and aquick start signal 108. The quickstart adjustment signal 96 is coupled to control an adjustable delay element 92. The output of delay element 92 and theHSON signal 52 are coupled to AND gate 94 that generates thequick start signal 108, as shown. - A
logic gate 132 receives theHSON signal 52 and thequick start signal 108 and provides an output signal to start thefirst delay element 130 a on the falling edge of the quick start signal. The output of thefirst delay element 130 a is coupled to an input of thesecond delay element 130 b and the output of thesecond delay element 130 b is coupled to an input of thethird delay element 130 c, as shown. In anexample driver 104, thefirst delay element 130 a provides a 12 ns delay, thesecond delay element 130 b provides a 7 ns delay, and thethird delay element 130 c provides a 7 ns delay.Logic gate 128 a receives theHSON signal 52 and logic gates 128 b-128 d receive the outputs of the three delay elements 130 a-130 c, respectively, and each gate 128 a-128 d additionally receives thequick start signal 108. This configuration results in the first driver stage (102 a) being on when theHSON signal 52 is active. - In the illustrated embodiment, the driver
control signal generator 152 is implemented on a “low side” of theregulator 100 in the sense that its signal levels are logic signal levels; whereas, thedriver stage circuitry 150 is on a “high side” of the regulator, with the SW node signal level ranging from VIN 13 (FIG. 2 ) to ground. TheBOOT signal 114 may be provided by a bootstrap capacitor (not shown) connected between BOOT and theswitch node SW 11 in order to maintain the required gate drive voltage. The bootstrap capacitor may be charged when theSW node 11 approaches ground, such as with a diode connected to a voltage source. Level shifters 126 a-126 d are coupled between respective outputs of OR gates 128 a-128 d and thedriver stage circuitry 150 in order to translate the logic level signals associated with the OR gates 128 a-128 d to high side signal levels for coupling to thedriver stage circuitry 150. - It will be appreciated that while, in the illustrated
multi-stage driver 104, the drivercontrol signal generator 152 is implemented on the low side, this signal generating circuitry could alternatively be implemented on the high side, thereby reducing the required number of level shifters 126 a-126 d. The decision to implement thesignal generator 152 on the low side versus the high side impacts silicon area, ESD protection, and other factors depending on the process. For example implementing delay cells 130 a-130 c on the high side can result in a larger physical circuit for each delay cell, but to implement them on the low side requires more level shifters, which also requires more physical area. - In addition to the plurality of driver stages 102 a-102 d, the
driver stage circuitry 150 includespre-driver buffer stages 124 a-124 d and 122 a-122 d. The buffer stages are sized to achieve a predetermined gate drive level for the driver stages 102 a-102 d and each buffer stage may have the same or different drive capability. It will be appreciated that additional or fewer pre-driver buffer stages may be provided. Here, the buffers 122 a-122 d and 124 a-124 d are provided in the form of inverters. - One of the level shifted signals, here a signal from
level shifter 126 a that corresponds to the delayed signal 140 a that controls thefirst driver stage 102 a, is coupled to abuffer inverter 120 for further coupling to abuffer inverter 118 and to a gate terminal of aNMOS FET 110. TheNMOS FET 110 has a drain terminal coupled to thegate terminal 112 of the high side switch 12 (FIG. 2 ) and a source terminal coupled to the switch node SW 11 (FIG. 2 ). In operation, theNMOS transistor 110 is off when any of the PMOS driver transistors 102 a-102 d is on. - Referring also to the illustrative waveforms of
FIG. 4 , operation of themulti-stage driver 104 will be explained.FIG. 4 shows anexample HSON signal 52 andquick start signal 108. Also shown inFIG. 4 are the driver control signals 140 a-140 c, the gatecurrent Igate 160 of the high side switch 12 (FIG. 2 ), the gate to source voltage,HSgate 164 of theswitch 12, and the switchnode voltage SW 11. - Each of the illustrated signals is shown in connection with driving three different high side switches 12, each having a different size and therefore different input capacitance. A
first portion 204 of the waveforms illustrates the respective signals when themulti-stage driver circuit 104 is coupled to a relatively smallhigh side FET 12 with a relatively small input capacitance, asecond portion 206 of the waveforms illustrates the signals when themulti-stage driver circuit 104 is coupled to a mediumsized FET 12, and athird portion 208 of the waveforms illustrates the signals when themulti-stage driver circuit 104 is coupled to a relativelylarge FET 12. - Details of the quick start controller 106 (
FIG. 2 ) are described below in connection with quick start control and calibration circuit 510 ofFIG. 11 . Suffice it to say here that under certain operating conditions (e.g., when driving atransistor 12 of a certain size), the quick start signal 108 (via OR gates 128 a-128 d) turns on all of the gate drive stages 102 a-102 d initially, at the rising edge of theHSON signal 52, for a quick start interval that ends when the gate to source voltage,HSgate 164 is at or near (but not above) the Miller plateau as is labeled 252 for the mediumsized FET 206 and 254 for thelarge FET 208. Thequick start signal 108 is provided in the form of a pulse that commences in response to a transition of theHSON signal 52 and ends when the switchnode voltage SW 11 starts to rise. In this way, the end of the quick start pulse or interval can be considered to provide an indicator of the start of the slew time interval and/or the transistor being at or near its Miller plateau. Use of thequick start signal 108 to quickly charge the transistor to a point at or near its Miller plateau can reduce the total duration of the transition of the gate to source voltage Vgs of the transistors. - Following the quick start interval, all but the
first driver stage 102 a (driven by the HSON signal) is turned off (i.e., driver control signal 140 a stays high rather than going low to turn offdriver stage 102 a). Stated differently, thequick start pulse 108 turns on all of the drive stages 102 a-102 d with its rising edge and turns off all but thefirst drive stage 102 a with its falling edge. Thereafter, the second through fourth driver stages are turned on by respective driver control signals 140 b-140 d according to their respective delays, as shown. - The
first portion 204 of the waveforms ofFIG. 4 corresponds to driving ahigh side FET 12 that is so small that it does not receive aquick start pulse 108. Here, at the rising edge of theHSON signal 52, the first driver control signal 140 a transitions to a high state to turn on thefirst driver stage 102 a. A first delay time later, here 12 ns later as established bydelay element 130 a, the seconddriver control signal 140 b goes high to turn on thesecond driver stage 102 b. Subsequently, after a second delay time of 7 ns as established bydelay element 130 b, the thirddriver control signal 140 c goes high to turn on thethird driver stage 102 c and finally, after a third delay time of 7 ns as established bydelay element 130 c, the fourthdriver control signal 140 d goes high to turn on thefourth driver stage 102 d. - As is apparent from consideration of the
waveform portions 204 associated with driving thesmallest FET sized FET largest FET 12, different numbers of driver stages 102 a-102 d may be turned on during the slew time interval depending on the capacitance of the driver FET. For example, in the case of the smallest FET, only thefirst driver stage 102 a is required to charge theSW node 11 to the VIN voltage level (i.e., only onedriver stage 102 a is on during the slew time interval). In the case of the medium sized FET, the first and second driver stages 102 a, 102 b are required in order to charge theSW node 11 to the VIN voltage level (i.e., twodriver stages node voltage SW 11 would look like if only a single driver stage (e.g., 102 a) were on during the slew time interval. And in the case of the largest FET, threedriver stages SW node 11 to the VIN voltage level (i.e., threedriver stages node voltage SW 11 would look like if only the first and second driver stages 102 a, 102 b were on during the slew time interval and a dotted line waveform labeled 258 illustrates what the switchnode voltage SW 11 would look like if only thefirst driver stage 102 a were on during the slew time interval. - It will be appreciated that the delay elements 130 a-130 c may provide the same delays or different delays. Various considerations may be used to select the delays provided by the delay elements in order to ensure that a predetermined slew time interval is achieved for all possible sizes of driven FETs. As one example, the delays can be selected to achieve the predetermined slew time interval for the smallest driven FET with only a
first driver transistor 102 a on and to require some additional number of driver transistors to be on to achieve the same predetermined slew time interval for the largest driven FET. In the examplemulti-stage driver 104, thefirst delay element 130 a establishes a longer delay (i.e., 12 ns) than the second andthird delay elements quick start signal 108 is present or not. Preferably the quick start interval (when thequick start signal 108 is active) is not present during the Miller plateau. - As is apparent from the driver control signals 140 a-140 d, in the illustrated embodiment, each stage is sequentially turned on such that a subsequent stage is turned on while the stage(s) that were previously on remain on in order to thereby achieve a total driver impedance equal to the parallel combination of the impedances of all of the active driver stages. It will be appreciated that other sequencing schemes are also possible. Further, while the example
multi-stage driver 104 includes fourdriver stages 102 a . . . 102 d, other numbers of stages are possible. - The impedance of the driver transistors 102 a-102 d may be the same as each other or may be different in order to achieve a desired scaling of the total parallel impedance as driver stages are turned on. In an embodiment, the impedance of
driver stage 102 a is 30Ω, the impedance ofdriver stage 102 b is 30Ω, the impedance ofdriver stage 102 c is 15Ω and the impedance ofdriver stage 102 d is 5Ω. With this arrangement, during a first time interval, when only thefirst driver stage 102 a is on, the total driver impedance is 30Ω, during a second time interval when both thefirst driver stage 102 a and thesecond driver stage 102 b are on, the total driver impedance is 15Ω, during a third time interval when driver stages 102 a, 102 b, and 102 c are on, the total driver impedance is 7.5Ω, and during a fourth time interval when all four driver stages 102 a-102 d are on, the total driver impedance is 3Ω. - Referring again to
FIG. 3 , multiple driver stages 102 a-102 d are only used for the pull up devices (i.e., the PMOS FETs) of thedriver circuit 104. Additionally or alternatively, the multi-stage driver approach can be applied to the pull down gate drive (i.e., NMOS FET 110) or to the low side switch (i.e., FET 14), which can be required by different switcher topologies, such as an inverted buck LED driver topology where the inductor current flows in the opposing direction of the illustrated embodiment. The multi-stage driver approach may also be desirable on the low side of an H-Bridge buck-boost regulator constructed with two switch nodes SW, with a high side multi-stage driver on the input switch node SW and a low side multi-stage driver on the output switch node SW, as will be described in connection withFIG. 16 . In the illustrated embodiment, the pull up devices are focused on because controlling the slew rate of the SW rising edge can be more important than the falling edge. Since, during this transition, current is being sourced from the input supply VIN, which has a relatively large current loop, the slew rate of the rising edge of theswitch voltage SW 11 can have more of an impact on EMI/EMC performance. In contrast, for theswitch node SW 11 falling edge, the load current is being transitioned to ground, which will have tighter associated current loops and therefore less of an impact on EMU/EMC performance. Other embodiments may have need for the multi-stage driver approach applied to the pull down gate drive. - In addition to quickly charging the
switch 12gate 112 to a point at or near its Miller plateau, thequick start signal 108 can also be used as a mechanism for initiating operation of themulti-stage driver 104 by operation of gate 132 (FIG. 3 ). It will be appreciated that this functionality can be achieved with alternative circuits and techniques, such as an adaptive quick start period, a pre-programmable quick start period, or other methods. - Furthermore, the multi-stage driver can be implemented without a quick start interval. One such alternative regulator 170 that does not include a quick start interval to more quickly charge the
FET 12 to its Miller plateau is shown inFIG. 5 . Regulator 170 differs from the regulator 100 (FIG. 2 ) in that thequick start controller 106 is replaced with a switch node monitor 114 that generates aswRising signal 116 and themulti-stage driver 104 is replaced with an alternativemulti-stage driver circuit 180 that responds to theswRising signal 116 rather than to a quick start adjustment signal. In this implementation the falling edge ofquick start signal 108 is analogous to the rising edge of theswRising signal 116. However, the falling edge ofquick start signal 108 occurs before the start of the Miller plateau and the rising edge of theswRising signal 116 occurs after the Miller plateau. - The
swRising signal 116 is used to initiate operation of themulti-stage driver 104. More particularly, theswRising signal 116 provides an indicator of the start of the slew time interval. To this end, switch node monitor 114 monitors the voltage at theswitch node SW 11 to determine when theswitch 12 has reached its Miller plateau and may include a comparator to compare the switchnode voltage SW 11 to a reference voltage selected so that a transition of the swRising signal 116 (e.g., a transition to a logic high level) coincides generally to theswitch 12 reaching its Miller plateau as can be determined by monitoring when the switch node voltage starts to rise. Thus, theswRising signal 116 is at a first logic level when the switchnode voltage SW 11 is greater than the reference voltage and is at a second logic level when the switchnode voltage SW 11 is less than the reference voltage. The switch node monitor comparator may include hysteresis. In some embodiments, the reference voltage may be user programmable or otherwise programmable or adjustable. In order to better reject variations in the rise time of the switchnode voltage SW 11, it may be desirable to provide the reference voltage at a relatively low level, such as on the order of 1 volt for a 12 volt supply, or more generally, on the order of approximately 10% of the supply voltage VIN. It is also desirable for the delay between theSW node voltage 11 rising edge and theswRising signal 116 to be as small as possible. The swRising signal could also be generated by monitoring the HSgate signal with a threshold configured below the Miller plateau; however, the Miller plateau is often an unknown voltage. - Referring also to
FIG. 6 , themulti-stage driver circuit 180 includes a drivercontrol signal generator 190 anddriver stage circuitry 150. The drivercontrol signal generator 190 includes an AND gate 144 that receives theHSON signal 52 and theswRising signal 116, as shown. Thus, the output of gate 144 transitions to a high level upon a rising edge of theswRising signal 116 and transitions to a low level once theHSON signal 52 transitions to a low level to indicate that the Miller plateau has been reached. The drivercontrol signal generator 190 includes three delay elements, here elements 146 a-146 c, each providing a respective delay. In an embodiment, delay element 146 a establishes a delay of 5 ns, delay element 146 b establishes a delay of 7 ns, and delay element 146 c establishes a delay of 7 ns. With this arrangement, a first control signal 140 a turns on and off thefirst driver stage 102 a according to the timing of theHSON signal 52, asecond control signal 140 b turns on thesecond driver stage 102 b 5 ns after theswRising signal 116 rising edge (since the delay element 146 a is triggered by theswRising signal 116 going high), athird control signal 140 c turns on thethird driver stage 102 c 7 ns after thesecond driver stage 102 b is turned on, and afourth control signal 140 d turns on thefourth driver stage 102 d 7 ns after thethird driver stage 102 c is turned on. In this way, theswRising signal 116 initiates operation of themulti-stage driver 180 in the same fashion as the quick start signal 108 (FIG. 3 ) such that the delayed driver control signals commence at a controlled time following a transition of theswRising signal 116 that indicates the beginning of the slew time interval, but here without the initial quick start charging implemented by the quick start signal 108 (FIG. 3 ). Here again, it will be appreciated that the delays established by delay elements 146 a-146 c can be readily modified. - Referring to
FIG. 7 , aswitching regulator 200 implementing a dead time calibration includescontrol circuit 48 andregulator circuitry 90. Theregulator 200 further includes adriver circuit 210 that differs from the driver circuit 80 (FIG. 2 ) in that the fixeddelay element 45 is replaced with anadjustable delay element 212. A dead time control and calibration circuit, or simply control circuit, orcalibrator 240 includes circuitry and implements techniques for optimizing the dead time between turning off thelow side switch 14 and turning on thehigh side switch 12 by adjusting the delay established by thedelay element 212. Thedead time calibrator 240 sets a dead time adjustment to an initial value and changes the dead time adjustment while monitoring a delay between the switchnode voltage SW 11 rising and thelow side switch 14 turning off to determine when to stop changing the dead time adjustment based on when the delay stops changing, as will be described. It will be appreciated that while the dead time calibration circuit is described in connection with certain embodiments (FIGS. 7 and 11 for example) to optimize the dead time between turning off thelow side switch 14 and turning on the high side switch, in other embodiments (FIG. 16 for example) dead time calibration optimizes the dead time between turning off the high side switch and turning on the low side switch in which case the dead time calibrator sets a dead time adjustment to an initial value and changes the dead time adjustment while monitoring a delay between the switch node voltage falling and the high side switch turning off to determined when to stop changing the dead time adjustment based on when the delay stops changing. - The
switch monitor 114, as described above in connection withFIG. 5 , compares the switchnode voltage SW 11 to a reference voltage to generate theswRising signal 116. Thus, theswRising signal 116 is at a first logic level when the switchnode voltage SW 11 is greater than a reference voltage and is at a second logic level when the switch node voltage is less than the reference voltage. In an embodiment, the reference voltage is relatively low, such as on the order of 1 volt for a 12 voltsupply voltage VIN 13. In this way, theswRising signal 116 provides an indication of theswitch node voltage 11 rising. - A
delay comparator 214 receives theswRising signal 116, anLSgate signal 56 coupled to the gate terminal of thelow side switch 14, and asampleDead signal 230 and compares theswRising signal 116 to theLSgate signal 56 to determine a present delay between thevoltages example delay comparator 214 ofFIG. 9 . Outputs of thedelay comparator 214 include adeadLonger signal 226 that, when high for example, indicates that the present delay is longer than the past delay and adeadShorter signal 228 that, when low for example, indicates that the present delay is shorter than the past delay. Additionally, delaycomparator 214 drives both thedeadLonger signal 226 and thedeadShorter signal 228 low, for example, indicating that the present delay is substantially equal to the past delay. - A dead time
calibration state machine 244 responds to thedeadLonger signal 226 and to thedeadShorter signal 228 and implements a calibration routine whereby, the dead time adjustment is either decremented or incremented or unchanged. To this end, thestate machine 244 generates an incDead signal 220 to increment the dead time adjustment and adecDead signal 222 to decrement the dead time adjustment via adead register 216. - The delay of the
delay element 212 is controlled by an address 218 (i.e., by a dead time adjustment 218) from thedead register 216 and, in an embodiment, can have a width on the order of 5 bits and a LSB weight of 1-3 ns for example. Thedead register 216 generates and stores the address 218 with an up/down counter controlled by the incDead signal 220 and thedecDead signal 222. AresetDead signal 224 resets theregister 216 to set theadjustable delay element 212 to an initial, minimum delay value, resulting in the most positive (i.e., maximum) dead time for the system. - Note that when decrementing the dead time adjustment with the
decDead signal 222, the delay of theadjustable delay element 212 actually increases. Conversely, when incrementing the dead time adjustment with the incDead signal 220, the delay of theadjustable delay element 212 decreases. This inversion is the result ofadjustable delay element 212 being coupled to thelow side switch 14 to delay theLSgate signal 56 relative to theHSON signal 52. - As will be explained further in connection with the flow diagram of
FIG. 10 , thedead time calibrator 240 initially sets the dead time adjustment to an initial value (here, to a minimum adjustment value) corresponding to an initial, maximum dead time. Thecalibrator 240 slowly attempts to decrement the dead time by incrementing the dead time adjustment by a small step, such as on the order of between 1 and 3 nanoseconds, following which the dead time is monitored to determine if a change in the delay between theLSgate signal 56 and the swRising signal 116 (i.e., a change in the dead time) has occurred. If a change in the dead time occurs, then the dead time adjustment is again incremented and the process is repeated. Once the dead time no longer changes in response to an increment of the dead time adjustment, the dead time is determined to be negative and therefore the previous step was at or near an optimal value. Stated differently, thedead time calibrator 240 operates on the principle that the optimal dead time occurs slightly before when the relative delay between the switchnode voltage SW 11 and the low side switchcontrol signal LSgate 56 no longer changes in response to a change (e.g., an increment) of the dead time adjustment. It will be appreciated that in embodiments (e.g.,FIG. 16 ) in which the dead time calibration controls the dead time between the high side switch turning off and the low side switch turning on, the optimal dead time occurs slightly before when the relative delay between the switch node voltage and the high side switch control signal no longer changes in response to a change of the dead time adjustment. The reason that no further change in the dead time is indicative of being at or near the transition from a positive dead time to a negative dead time is because this condition occurs when the dead time is negative. In many applications, it is desirable to have a slight positive dead time. - Referring also to
FIG. 8 , example signal waveforms associated with theregulator 200 ofFIG. 7 include the switchnode voltage SW 11, theLSgate signal 56, the high side switch drain to sourcecurrent Ids 314, and the high side switch gate to source voltage, or control signalHSgate 316. Each of the illustrated signals is shown in connection with various dead times. Afirst portion 304 of the waveforms illustrates the respective signals when theregulator 200 is operating with several different dead times (with the left most waveforms corresponding to the most negative dead time), asecond portion 306 of the waveforms illustrates the signals when theregulator 200 is operating with a first, positive dead time, athird portion 308 of the waveforms illustrates the signals when theregulator 200 is operating with a second, optimal dead time, and afourth portion 310 of the waveforms illustrates the signals when theregulator 200 is operating with a third, negative dead time. - As the dead time is reduced, the peak drain current 314 decreases with each step because the body diode of the
low side switch 14 has less time for charge to be built up. The conduction of thelow side switch 14 body diode results from the switchnode voltage SW 11 going negative. When the dead time becomes negative (310), the peak of drain current 314 significantly increases due to thehigh side switch 12 and thelow side switch 14 simultaneously conducting (shoot through current). Also as illustrated in thewaveform portions 304, when the dead time crosses from positive to negative, the switch node slew rate increases dramatically. This increase in slew rate is due to excess gate to sourcevoltage 316 on thehigh side FET 12 when thelow side switch 14 and thehigh side switch 12 conduct simultaneously. As will be explained in connection withFIG. 10 , the dead time calibration routine selects as the optimal dead time a dead time that is one or two steps from the point in time (i.e., the dead time adjustment 218) where relative delay between the low side switch turning off and the switch node voltage rising no longer changes. - Referring also to
FIG. 9 , anexample delay comparator 214 is shown to receive theLSgate signal 56 and theswRising signal 116 and to provide thedeadLonger signal 226 and thedeadShorter signal 228. Thedelay comparator 214 is additionally responsive to asampleDead signal 230 from the dead time calibration state machine 244 (FIG. 7 ). Thedelay comparator 214 measures and stores the present delay between theLSgate signal 56 and theswRising signal 116 when thesampleDead signal 230 is high and otherwise determines the relative delay (i.e., compares the present delay to a stored past delay). Suffice it to say here that thedeadLonger signal 226 indicates when the present delay is longer than the past delay and thedeadShorter signal 228 indicates when the present delay is shorter than the past delay. - The
LSgate signal 56 is coupled to an inverter 705 and to adelay line 700 including a plurality of delay cells 704 a-704 y configured to provide a plurality of delay signals 728 to unit cells 702 a-702 y. In order to ensure reliable detection, theadjustable delay element 212 and thedelay comparator 214 are designed using similar delay cells so that shifts due to process, temperature, or bias conditions will tend to cancel. Thus, thedelay comparator 214 has a resolution on the order of the least significant bit (LSB) weight of the adjustable delay element 212 (FIG. 7 ). If the comparison is less than one LSB of theadjustable delay element 212, then thedelay comparator 214 will cause both thedeadLonger signal 226 and thedeadShorter signal 228 to be low, thereby indicating that there has been no change in the delay (i.e., the past delay is substantially the same as the present delay). As an example, each delay cell 704 a-704 y may have a delay of about one-half the LSB delay step of the adjustable delay element 212 (FIG. 7 ). With this arrangement, thedelay comparator 214 is able to reject small deviations in the delay comparison that are not attributable to the LSB delay step of theadjustable delay element 212. - It will be appreciated that the delay cells 704 a-704 y can establish the same or different delays as each other. Using different magnitudes of delays can allow the
delay comparator 214 to measure a wider range of delays than otherwise possible, since the total sum of the delay established by thedelay line 700 corresponds to the maximum delay that thedelay comparator 214 can measure. In one example, a first plurality of the delay cells 704 a-704 y can provide a first predetermined delay and a second plurality of the delay cells can provide a second predetermined delay that is shorter than the first predetermined delay. With this type of configuration, coarser steps can be used to get an approximate measurement of the delay and then finer delay steps can be used to “tune” the measurement to the actual delay with high resolution. - Each unit cell 702 a-702 y includes a
first register 714, as may be implemented with a latch, to store the present delay and provide apresent delay signal 726 and asecond register 716, as also may be implemented with a latch, to store the past delay and provide apast delay signal 728. Thepresent delay register 714 is strobed every period that theswitch node 11 is active (i.e. during the slew time interval) by astrbDead signal 730. ThestrbDead signal 730 is generated in response to theLSgate signal 56, thereby latching the register on the rising edge ofswRising signal 116, and theswRising signal 116 and in one example, corresponds to the dead time when neither transistor is on. The duration of thestrbDead signal 730 defines the interval during which the present delay between theLSgate signal 56 and theswRising signal 116 is measured and stored. - The sampleDead signal 230 (
FIG. 7 ) is buffered by abuffer 708 to provide astrbReg signal 732 that enables thepast register 716 to transfer the present delay from thepresent delay register 714 to thepast delay register 716. This transfer may occur when thestrbDead signal 730 is low (i.e., when thepresent delay register 714 is disabled). -
Logic gates present delay signal 726 and thepast delay signal 728, as shown. The output signals 734 fromgates 718 of the plurality of unit cells 702 a-702 y are coupled throughfurther gates 722, 710 to generate thedeadShorter signal 228 for coupling to the dead time calibration state machine 244 (FIG. 7 ). Similarly, the output signals 736 from thegates 720 from the plurality of unit cells 702 a-702 y are coupled throughfurther gates deadLonger signal 226 for coupling to the dead time calibration state machine 244 (FIG. 7 ). - In the example embodiment, for the
deadShorter signal 228 or for thedeadLonger signal 226 to be at a logic high level, at least two consecutive unit cells 702 a-702 y must have a mismatch between thepast delay signal 728 and thepresent delay signal 726. More particularly, when thepresent delay signal 726 is low and thepast delay signal 728 is high, then theoutput signal 734 will be at a logic high, and if twoconsecutive outputs 734 are high,logic gate 722 will pass a logic high signal through OR gate 710 forcing thedeadShorter signal 228 to be at a logic high level. Similarly, when thepresent delay signal 726 is high and thepast delay signal 728 is low, then theoutput signal 736 will be at a logic high, and if twoconsecutive outputs 736 are high,logic gate 724 will pass a logic high signal through ORgate 712 forcing thedeadLonger signal 226 to be at a logic high level. - It will be appreciated that other techniques can be used to implement the
delay comparator 214. However, the described digital approach may reject process variations and mismatches better than some other techniques, such as analog techniques. - Referring also to
FIG. 10 , a flow diagram illustrates atechnique 400 implemented by thedead time calibrator 240 ofFIG. 7 . The rectangular elements (typified by element 402) are herein denoted “processing blocks” and the diamond-shaped elements (typified by element 404) are herein denoted “decision blocks” and either or both may represent computer software instructions or groups of instructions. It should be noted that the flow diagram ofFIG. 10 (and other flow diagrams herein) represent exemplary embodiments of designs disclosed herein and variations in such embodiments, which generally follow the processes outlined, are considered to be within the scope of the concepts, systems and techniques described and claimed herein. Some or all of the blocks may represent operations performed by functionally equivalent circuits. Also, some blocks may be manually performed while other blocks may be performed by machine. The flow diagrams do not depict the syntax of any particular programming language. Rather, the flow diagrams illustrate the information one of ordinary skill in the art requires to fabricate circuits and/or to generate computer software to perform the processing required of the particular apparatus. It should be noted that many routine program elements, such as initialization of loops and variables and the use of temporary variables are not shown. It will be appreciated by those of ordinary skill in the art that unless otherwise indicated herein, the particular sequence described is illustrative only and in instances can be varied without departing from the spirit of the concepts described and/or claimed herein. Thus, unless otherwise stated, the processes described below are unordered meaning that, when possible, the actions shown in the diagrams can be performed in any convenient or desirable order, including simultaneously. - The dead time is initially set to a maximum value in
block 401 by theresetDead signal 224. Thus, thedead register 216 may be set to a value corresponding to a minimum delay bydelay element 212 in order to achieve a maximum dead time target, since for a delay coupled to the low side switch14, an attempt to decrement the dead time is equivalent to increasing the delay provided bydelay element 212. Thereafter, a deadtime decrementing phase 440 is entered during which it is attempted to decrement the dead time atblock 402, following which the relative delay between the switchnode voltage SW 11 and theLSgate signal 56 is determined atdecision block 404. Before the dead time is attempted to be decremented (and herein, before each time that the dead time is attempted to be decremented or incremented as also occurs atblocks present delay register 714 to the past delay register 716 (FIG. 9 ) in response to strbReg signal 732 (FIG. 9 ) as indicated by the “sampleDead” statement inblocks - If it is determined at
decision block 404 that the present delay is longer than the past delay (e.g., as may be indicated by the deadLonger signal 226), then the dead time is again attempted to be decremented atblock 402. Similarly, if it is determined that the present delay is shorter than the past delay (e.g., as may be indicated by the deadShorter signal 228), then the dead time is again attempted to be decremented atblock 402 following a wait period atblock 406. The present delay being shorter than the past delay indicates that the dead time is positive. The wait period block 406 may be on the order of sixty switching cycles (set by the bandwidth of the regulation loop) to ensure that the calibration routine does not respond to transients. In the unlikely event that a line or load transient occurs and the delay comparison atblock 404 indicates an erroneous longer present dead time than past dead time, then an attempt will be made immediately to decrement the dead time atblock 402 to ensure that proper operation is maintained. It will be appreciated that the delay compareblock 404 and waitblock 406 alternatively could be implemented by averaging the delay comparison over several cycles and proceeding to attempt to decrement the dead time if a majority of the cycles yields a shorter relative delay. - If it is determined at
block 404 that the delay has not changed (i.e., the present delay is substantially equal to the past delay as indicated by both thedeadLonger signal 226 and thedeadShorter signal 228 being low for example), then a dead time tuning andvalidation phase 442 is entered. More particularly, validation is performed at decision block 418 to validate that an optimal dead time has been achieved. No change in the relative delay following the deadtime decrement block 402 indicates that the dead time has crossed from being a positive dead time to a negative dead time. Once the dead time is negative, further attempts to decrement the dead time will not cause a change in the result of the delay comparator 214 (FIG. 7 ), since thelow side switch 14 solely determines when theswitch node SW 11 is released. Any further attempts to decrement the dead time at this point will increase the shoot through current and the switch node slew rate, due to overcharging the high sidegate voltage HSgate 316 as shown inFIG. 8 . Given the significant shoot through current, the delay element 212 (FIG. 7 ) is immediately decremented (i.e., the incDead signal 220 is incremented inblock 414 without a wait state) in order to impact the efficiency as little as possible. - The dead time corresponding to the point at or very near to when the dead time crosses from being positive to negative can be considered an optimal dead time. Before validating the dead time at block 418, the dead time may be incremented at
block 414 as may be desirable to back off slightly from the dead time setting at which a positive dead time just becomes a negative dead time in order to minimize the efficiency loss associated with a negative dead time. Following await state 416, it is determined at validation decision block 418 whether the relative delay is still unchanged. Thewait state 416 causes the validation 418 to be performed over many cycles, such as on the order of 60 cycles, to ensure that transients do not affect the determination. - If it is determined in decision block 418 that the relative delay has changed, then the optimal dead time has not been found. More particularly, if the present delay is determined to be longer than the past delay (430), then the calibrator returns to block 402 to again attempt to decrement the dead time. Alternatively, if the present delay is determined to be shorter than the past delay (428), then the calibrator attempts to increase the dead time (and additionally captures the relative delay) at
block 408 and again determines atdecision block 411 whether the relative delay has changed. If it is determined that the present delay is longer than the past delay, it is again attempted to decrement the dead time atblock 402. If however it is determined that the present delay is either shorter than the past delay or that there has been no change in the delay even following the increment inblock 408 due to a negative dead time, then the dead time is again incremented atblock 408. Thisadditional comparison block 411 ensures that a positive dead time is achieved so that thedecrementing phase 440 can find the optimal dead time. - If it is determined in validation block 418 that the relative delay has not changed, then the validation has passed (432). In this case, a further increment of the dead time may be made in
block 420 in order to further increase the efficiency, following which a dead timecalibration termination phase 446 may be entered. More particularly, the relative delay is captured inblock 421 by thesampleDead signal 230 going high so that subsequent changes in the dead time can be detected. - Thereafter, the system idles at
block 412 and continuously checks the relative delay atdecision block 410 in order to determine whether a change in the system operating conditions results in the optimal dead time changing. A determination atdecision block 410 that the present delay is shorter than the past delay (434) can indicate that the threshold voltage of theswitches block 408 until the present dead time becomes longer than the past dead time. If it is determined atdecision block 410 that the present delay is longer than the past delay, then the dead time has become more positive due to changing operating conditions and thedecrementing phase 440 recommences, as shown. - The dead
time calibration method 400 presumes that theswitch node 11 is switching and that the regulator operating conditions have not dramatically changed. In order to ensure proper operation, aresetCal signal 234 and aholdCal signal 232 are provided to the state machine (FIG. 7 ), as may be from an external control unit or processor. TheresetCal signal 234 is forced high during system events that significantly change the operating point of theswitch node SW 11, such as start up of the regulator or fault events. When theresetCal signal 234 is high, thestate machine 244 is forced into and held at aninitial state 401 which also forces theresetDead signal 224 high to reset thedead register 216 to the initial, most positive dead time setting. For less significant events, or simply when operation of theswitch node SW 11 is held off for a brief period of time, theholdCal signal 232 is used. When theholdCal signal 232 is high, the contents of thedead register 216 remain unchanged and thestate machine 244 is held at a wait or comparison state (e.g., states 404, 406, 416, 418, 411, 421, 412, or 410). Once theholdCal signal 232 goes low, theprocess 400 resumes. - It will be appreciated that while the
dead time calibrator 240 is described as adjusting only the delay between the high sidecontrol signal HSON 52 and the low sidecontrol signal LSgate 56, in some embodiments, it may be desirable to additionally or alternatively control a delay associated with driving thehigh side switch 12. As one example, the above circuitry and techniques can additionally include a further delay element between theHSON signal 52 and the gate terminal of thehigh side switch 12 with which the high side control signalHSgate 316 can be delayed relative to the low sidecontrol signal LSgate 56 under certain operating conditions, such as when thedelay element 212 sets the maximum initial dead time (i.e., corresponding to a minimum delay) or during a predetermined range of dead times including the maximum initial dead time. - Referring also to
FIG. 11 , aswitching regulator 500 that implements a quick start calibration in addition to dead time calibration includescontrol circuit 48,regulator circuitry 90, a driver circuit 504, dead time control and calibration circuit, orcalibrator 240, and a quick start control and calibration circuit, or simply control circuit, or calibrator. - The driver circuit 504 includes a
quick start driver 512 that is responsive to theHSON signal 52 and generates a high side control signalHSgate 508 for coupling to the gate terminal of thehigh side switch 12. Thedriver 512 is further responsive to a quick start adjustment adjQstart signal 522 from the quick start control circuit 510 and provides aquickStart signal 520 to the quick start control circuit 510, as shown. ThequickStart signal 520 may be the same as or similar to the quick start signal 108 (FIG. 3 ). Thus, thequick start signal 520 establishes an initial driver interval (referred to herein alternatively as the quick start interval) during which theswitch 12 is quickly charged to at or near its Miller plateau, without overshoot. ThequickStart signal 520 can be provided in the form of a pulse that commences in response to a transition of the low sidecontrol signal LSgate 56 and ends when the switchnode voltage SW 11 starts to rise. In this way, the end of the quick start signal, or pulse can be considered to provide an indicator of the start of the slew time interval. It will be appreciated that an optimal quick start interval will vary based on the threshold voltage and the gate capacitance of thehigh side switch 12. Because of this variation, the calibrator 510 includes circuitry and implements techniques to tailor the quick start interval to the switch parameters. - Quick start calibrator 510 includes a
state machine 514, aquick start comparator 516, and aQstart register 518. Thequick start comparator 516 monitors thequickStart signal 520 and thestate machine 514 adjusts the quickStart signal via theadjQstart signal 522 under certain conditions. In an embodiment, thequick start comparator 516 is configured to determine if a dead time (i.e., a time when neither thehigh side switch 12 nor thelow side switch 14 is on) resulting from thequickStart signal 520 is greater than an upper limit related to a maximum dead time target. Thequick start comparator 516 may additionally or alternatively monitor thequickStart signal 520 to determine if the quick start pulse terminates after a lower limit related to thehigh side switch 12 turning on (i.e., when theswitch node voltage 11 begins to rise). - An example
quick start comparator 516 is shown inFIG. 13 . Suffice it to say here that thecomparator 516 generates adeadLongError signal 530 that provides an indication of whether the dead time exceeds the upper limit and adeadShortError signal 532 that provides an indication of whether the quickStart signal pulse terminates after the lower limit. In one example, both thedeadLongError signal 530 and thedeadShortError signal 532 are low if the dead time is less than the upper limit and the quick start pulse terminates before the lower limit, thedeadLongError signal 530 is high and thedeadShortError signal 532 is low if the dead time is greater than the upper limit, and thedeadLongError signal 530 is low and thedeadShortError signal 532 is high if the quickStart pulse terminates after lower limit. - The
state machine 514 performs a quick start calibration routine to generate an incQstart signal 524, a decQstart signal 526, and aresetQstart signal 528, all of which are coupled to theQstart register 518. TheQstart register 518 generates theadjQstart signal 522 with an up/down counter in response to the incQstart signal 524, the decQstart signal 526, and/or theresetQstart signal 528 to control a delay in thedriver 512. In one example embodiment, theaddress bus 522 has a width of 5 bits and results in a LSB weight of 0.3 ns to 3 ns, for example. Anexample driver 512 is shown inFIG. 14 and described below. - Referring also to
FIG. 12 , example signal waveforms associated with theregulator 500 ofFIG. 11 include theswitch node voltage 11, the high side switch drain to sourcecurrent Ids 314, the high side control signalHSgate 508, the low sidecontrol signal LSgate 56, and thequickStart signal 520. Each of the illustrated signals is shown in connection with various operating conditions. Afirst portion 574 of the waveforms illustrates the respective signals when theregulator 500 is operating with several different quick start intervals decreasing from left to right, asecond portion 576 of the waveforms illustrates the signals when theregulator 500 is operating with a first quick start interval, and athird portion 578 of the waveforms illustrates the signals when theregulator 500 is operating with a second, optimal quick start interval, and afourth portion 580 of the waveforms illustrates the signals when theregulator 500 is operating with a third quick start interval. - Consideration of the waveforms illustrates that as the quick start interval increases, the gate to source
voltage 508 reaches the Miller plateau 588 in less time and the switchnode voltage SW 11 rises earlier.Waveform potion 576 illustrates an operating point when the quick start interval is considered too short, as can result in a dead time that is greater than a maximum dead time target (i.e., greater than an upper limit) necessary to achieve a relatively short switch on time target. This condition is flagged by the deadLongError signal 530 (FIG. 11 ).Waveform portion 578 illustrates an optimal quick start interval as is apparent by the quick start interval stopping just before theMiller plateau 592 is reached, thereby allowing a higher impedance gate drive to control the switch node as shown inFIG. 14 . Thelast portion 580 of the waveforms illustrates that when the quick start interval is too long, as can result when the quick start interval terminates after the switchnode voltage SW 11 begins to rise (i.e., when the quick start pulse terminates after the lower limit), the gate to sourcevoltage HSgate 508 can overshoot, resulting in a fast switch node rise time. This condition is flagged by the deadShortError signal 532 (FIG. 11 ). - Referring also to
FIG. 13 , an examplequick start comparator 516 includes afirst delay element 554 that is responsive to an inverted version of theISgate signal 56 provided by aninverter 550 to generate an input signal for alatch 558. Thedelay element 554 sets an upper limit for the regulator dead time and may be selected based on the minimum on time of theSW node 11, such as on the order of 10 ns for a 75 ns minimum on time. More particularly, the delay established bydelay element 554 may be selected to provide a relatively small permissible margin beyond a predetermined maximum dead time target. - A
second delay element 560 responsive to thequickStart signal 520 sets a lower limit for the quick start interval and provides an input signal to alatch 570. The delay established bydelay element 560 may be selected to provide a relatively small permissible margin beyond high side switch turning on (i.e., when the switchnode voltage SW 11 begins to rise), but during which the quick start pulse still may be active. Therefore,delay element 560 accounts for the inherent delay associated with translating the control signals on the low side to the output (i.e., to thegate terminal 112 of the high side switch) on the high side. -
Latches low side switch 14 turns off to a time when the switchnode voltage SW 11 begins to rise thereby latching the state on theswRising signal 116 rising edge, as is achieved in response to the strbDead signal 568 from alogic gate 564. The output oflatch 558 is thedeadLongError signal 530 that is indicative of the dead time being greater than the upper limit. The output oflatch 570 is thedeadShortError signal 532 that is indicative of thequickStart signal 520 being asserted outside of the strobe period (i.e., the quick start interval ending after the lower limit when the switch node voltage begins to rise). - Referring also to
FIG. 14 , an example quickstart driver circuit 512 includes a drivercontrol signal generator 640 anddriver stage circuitry 644. Thedriver stage circuitry 644 includes a plurality of driver stages 660 a-660 b, each having a control input responsive to a respective driver control signal 662 a-662 b generated by the drivercontrol signal generator 640 and an output coupled to the output of the other ones of the plurality of driver stages and to the control terminal (i.e., the gate terminal) of the high side switch 12 (FIG. 11 ). In the illustrated embodiment, each driver stage 660 a-660 b includes a driver transistor, such as in the form of the illustrated PMOS FETs, having a control input provided by its gate terminal and an output provided by its drain terminal. Thedriver transistors - Each driver control signal 662 a-662 b has an on time during which the respective driver transistor is on and an off time during which the respective driver transistor is off. At least one of the driver control signals 662 a-662 b has an on time that is controlled by the
quickStart signal 520 and at least another one of the driver control signals has an on time controlled by a combination of thequickStart signal 520 and theHSON signal 52. With this arrangement, during the quick start interval (i.e., when thequickStart signal 520 is active), both of the driver stages 660, 660 b are on; whereas, after the quick start interval, only one of the driver stages 660 a, 660 b is on. When more than one driver stage is turned on (during the quick start interval), the total impedance of the parallel driver stages is decreased as compared to when only one driver stage is on, in order to thereby allow thegate 112 of thehigh side switch 12 to charge more quickly than otherwise possible. Thus, in the illustrated embodiment, both of the driver stages 660 a, 660 b are on during the quick start interval and thereafter, only one of the driver stages, 660 b is on during the slew time interval of the switch 12 (as the switchnode voltage SW 11 rises). - The driver
control signal generator 640 includes anadjustable delay element 650 to generate thequickStart signal 520 in response to theHSON signal 52 and theadjQstart signal 522. Alogic gate 652 receives theHSON signal 52 and an output signal from theadjustable delay element 650 and generates thequickStart signal 520.Logic gate 654 a receives theHSON signal 52 and logic gates 654 a-654 b receive thequickStart signal 520, as shown. - Level shifters 664 a-664 b are coupled between respective outputs of OR gates 654 a-654 b and the
driver stage circuitry 644 in order to translate the logic level signals associated with the OR gates 654 a-654 b to high side signal levels for coupling to thedriver stage circuitry 644. As noted above in connection with other driver circuits, the drivercontrol signal generator 640 could alternatively be implemented on the high side, thereby reducing the number of required level shifters 664 a-664 b. - In addition to the plurality of driver stages 660 a-660 b, the
driver stage circuitry 644 includes pre-driver buffer stages 678 a-678 b, 680 a-680 b. The buffer stages are sized to achieve a predetermined gate drive level for the driver stages 660 a-660 b and each buffer stage may have the same or different drive capability. It will be appreciated that additional or fewer pre-driver buffer stages may be provided. Here, the buffers 678 a-678 b and 680 a-680 b are provided in the form of inverters. - One of the level shifted signals, here a signal from
level shifter 664 a that corresponds to the delayedsignal 662 a that controls thefirst driver stage 660 a, is coupled to abuffer inverter 672 for further coupling to abuffer inverter 674 and to a gate terminal of anNMOS FET 670. TheNMOS FET 670 has a drain terminal coupled to the gate terminal of the high side switch 12 (FIG. 11 ) and a source terminal coupled to the switch node SW 11 (FIG. 11 ). In operation, theNMOS transistor 670 is off when any of the PMOS driver transistors 660 a-660 b is on. - Referring also to
FIG. 15 , a flow diagram illustrates atechnique 600 implemented by the quick start calibrator 510 ofFIG. 11 . Atblock 601, theQstart register 518 is reset in response to theresetQstart signal 528. It is then determined atblock 602, such as by comparator 516 (FIG. 11 ), whether a dead time when neither thehigh side switch 12 nor thelow side switch 14 is on that occurs in response to thequickStart signal 520 is greater than an upper limit related to a maximum dead time target and whether the quick start signal pulse terminates after a lower limit related to the high side switch turning on. If neither of these conditions is true (i.e., if both the dead time is less than the upper limit and the quick start pulse terminates before the lower limit), then the quick start signal is considered to be within an acceptable range and the calibration routine proceeds to await state 604 in order to ensure that the result fromdecision block 602 was not due to a transient. Illustrative wait intervals implemented bywait block 604 may be on the order of 60 PWM cycles. The function of comparison by thequick start comparator 516 inblock 602 and thewait state 604 could alternatively be implemented by averaging the comparison over several cycles and proceeding forward with the comparison result that has the highest average. - After
wait state 604, the state machine 514 (FIG. 11 ) then rechecks the quickStart signal pulse width again to determine if it is within the comparison window (i.e., if the dead time is less than the upper limit and the quick start pulse terminates before the lower limit) withdecision block 610. If it is determined atdecision block 610 that thequick start signal 520 within the comparison window (632), then the quick start calibration is terminated atblock 614. If it is determined atdecision block 610 that the dead time is greater than the upper limit (as may be indicated by the deadLongError signal 530), then the quick start interval is incremented atblock 606 and processing continues as described above. If on the other hand, it is determined atdecision block 610 that the quick start pulse terminates after the lower limit (as may be indicated by signal deadShortError signal 532), then the quick start interval is decremented atblock 608 and processing continues as described above. - If in
block 602 it is determined that the dead time is greater than the upper maximum dead time target (622), theQstart register 518 is incremented by the incQstart signal 524 atblock 606. Once thequick start signal 522 is incremented, it is determined atblock 612 whether theQstart register 516 is at the upper limit. If theQstart register 516 is not at the upper limit (0), then await block 611 is entered, following which the process repeats beginning withblock 602. In this way, the quickStart signal pulse width is repeatedly incremented for as long as theQstart register 516 is not at the upper limit. If alternatively, it is determined atblock 612 that theQstart register 516 is at the upper limit (1), then the quick start calibration is deemed completed atblock 614 because no further quick start can be applied. - Thereafter, the quick start signal is again monitored at
block 620. If inblock 620, the quick start pulse is determined to terminate after the lower limit (634), then the process returns to block 608 to decrement the pulse width. If alternatively it is determined atblock 620 that the dead time is greater than the upper limit or that the both the dead time is less than the upper limit and the quick start pulse terminates before the lower limit (636), then the calibration routine is completed atblock 614. - If in
block 602 it is determined that quick start pulse terminates after the lower limit when the high side switch turns on, then theQstart register 518 is decremented by the decQstart signal 526 atblock 608. Once thequick start signal 522 is decremented, it is determined atblock 616 whether theQstart register 518 is at its lower limit. If thequick start register 518 is at its lower limit (1), then the quick start calibration is completed without further rechecking atblock 618. If the deadShortErrror persists even after theQstart register 518 is at its lower limit, then the calibration is “permanently” completed inblock 618. This scenario can occur, for example, in the presence of a negative load current which can cause the switchnode voltage SW 11 to rise immediately after thelow side switch 14 is turned off. If however theQstart register 518 is not at its lower limit (0), then the quick start calibration is deemed completed atblock 614, following which the quick start pulse termination is again rechecked atblock 620. In this way, the quick start calibration includes decrementing the quickStart signal pulse width in response to an active deadShortError signal (624) since this signal indicates that the end of the quick start interval is too close to the Miller plateau and could result in the switchnode voltage SW 11 rising too fast. If the deadShortErrror persists even after theQstart register 518 is at its lower limit, as done afterdecision block 602, then the calibration is “permanently” completed inblock 618. - A
QcalDone signal 536 and a recalQstart signal 538 (FIG. 11 ) may be coupled between the quick startcalibration state machine 514 and the dead time calibration state machine 244 (in embodiments including both calibrations). The quickstart calibration routine 600 must complete before the dead time calibration routine 400 (FIG. 10 ) starts. This is because the quick start calibration will significantly alter the dead time possibly making it impractical to run the dead time calibration routine at the same time as the quick start calibration routine. In addition, to make the quick start calibration as effective as possible, the calibration must be performed with the most positive dead time, which is when the dead register 216 (FIG. 11 ) is held in reset. Since the quick start calibration must be completed before the dead time calibration routine begins, theQcalDone signal 536 may be used to initiate the start of the dead time calibration routine. At any of the blocks where the quick start calibration is deemed completed (blocks 614, 618), theQcalDone signal 536 is asserted. In addition, if the dead time calibration routine cannot complete (e.g., because the dead time is at its limit in state 402), the quickstart calibration routine 600 is signaled to restarted with therecalQstart signal 538. - The quick start calibration (like the dead time calibration described above) requires that the
switch node 11 be switching and that the regulator operating conditions not dramatically change. In order to ensure proper operation, theresetCal signal 234 and theholdCal signal 232 are provided to the quick start state machine 514 (FIG. 11 ), as may be from an external control unit or processor. Here again, theresetCal signal 234 is forced high during system events that significantly change the operating point of theswitch node SW 11, such as start up of the regulator or fault events. When theresetCal signal 234 is high, thestate machine 514 is forced into and held at aninitial state 601 which also forces theresetDead signal 224 high to reset theQstart register 518. For less significant events, or simply when operation of theswitch node SW 11 is held off for a brief period of time, theholdCal signal 232 is used. When theholdCal signal 232 is high, the contents of the Qstart register 518 remain unchanged and thestate machine 514 is held at a wait or comparison state (e.g., states 602, 610, 620, 604, and 611). Once theholdCal signal 232 goes low, theprocess 600 resumes. - It will be appreciated that while the
quick start signal 522 is described as controlling the duration of the quick start period, this signal may alternatively or additionally be used to control the drive impendence. For this implementation, the quick start period would be a fixed duration, 10 ns for example, and theadjQstart signal 552 would control the drive strength used during the fixed quick start period. The adjustable drive strength may be implemented with multiple drivers using a variety of on/off combinations to generate a variety of drive strengths. - It will be appreciated that while the
quick start signal 520 is described in connection withFIG. 11 as only controlling a single switch (high side switch 12), the optimized quick start signal can also be used to drive other switches in the system. Referring also toFIG. 16 , an H-Bridgebuck boost regulator 700 implementing quick start calibration and a dead time calibration includescontrol circuit 48, regulator circuitry 702, an input driver circuit 504 a, an output driver circuit 504 b, a dead time and quickstart calibration circuit 704 a to control the input driver circuit 504 a and a dead time and quickstart calibration circuit 704 b to control the output driver circuit 504 b. The H-bridge regulator circuitry 702 includes an inputhigh side switch 12 a, an input low side switch 14 a, an outputhigh side switch 12 b, and an output low side switch 14 b, coupled toinductor 16 as shown to generate a regulatedoutput voltage Vout 18. - The dead time and quick
start calibration circuit 704 a can be the same as or similar to a combination ofdead time calibrator 240 and quick start calibrator 510 (FIG. 11 ). The output driver 504 b differs from driver 504 a in that adelay element 212 b (that may be the same as or similar to delayelement 212 ofFIG. 11 ) is coupled to the gate of highside output switch 12 b throughbuffer 44 b and the quick start driver circuit 512 b (that may be the same as or similar to quickstart driver circuit 512 ofFIG. 11 ) is coupled to the gate of the low side output switch 14 b. The dead time and quickstart calibration circuit 704 b can thus, be the same as or similar to a combination ofdead time calibrator 240 and quick start calibrator 510 ofFIG. 11 , but is configured to controldelay element 212 b coupled to the highside output switch 12 b and to control quick start driver 512 b coupled to the low side output switch 14 b, as shown. Thus, whereas the dead time calibrator 240 (FIG. 11 ) is configured to calibrate the dead time between thelow side switch 14 turning off and thehigh side switch 12 turning on, thecalibrator 704 b is configured to calibrate the dead time between thehigh side switch 12 b turning off and the low side switch 14 b turning on. Accordingly, thecalibration circuit 704 a may be configured to monitor the delay between the low side switch 14 a turning off and the voltage at the switch node SWa rising in order to determine when to stop changing the dead time adjustment; whereas thecalibration circuit 704 b may be configured to monitor the delay between thehigh side switch 12 b turning off and the voltage at the switch node SWb falling in order to determine when to stop changing the dead time adjustment. - It will be appreciated that various alternatives may be implemented. For example, the quick start calibration implemented by
circuits circuit 704 a for example) may generate the adjQstart signal (signal 522 a for example) and that same adjQstart signal 522 a may be coupled to both the quick start driver 512 a and also to the quick start driver 512 b. The optimized quick start signal 522 a for example may also be used to drive the high side switch of a secondary high side switch (not shown) in the same system regulating a second load. Also, or alternatively, the dead time calibration implemented bycircuits delay element 212 a and also to thedelay element 212 b. - All references cited herein are hereby incorporated herein by reference in their entirety.
- Having described preferred embodiments, it will now become apparent to one of ordinary skill in the art that other embodiments incorporating their concepts may be used.
- It will be appreciated that the proposed circuitry and techniques can be applied to any linear or switching regulator topology including but not limited to Buck, Boost, Buck-Boost, SEPIC, Cúk, half-bridge, full bridge, and linear regulators utilizing with any type of control loop including current mode control, voltage mode control, constant on time control, constant off time control, or any other analog and/or digital control scheme. In addition, the proposed circuitry and techniques can be applied to regulators that regulate current, voltage, power, or other parameters. The circuitry and techniques described herein can be implemented using hardware, software, and/or firmware in a digital and/or analog fashion. Thus, it will be appreciated that certain terms used herein, such as controller, processor, control circuit, state machine, can be implemented in any suitable fashion and are not intended to require any particular implementation methodology.
- It is felt therefore that these embodiments should not be limited to disclosed embodiments, but rather should be limited only by the spirit and scope of the appended claims.
Claims (38)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US14/798,903 US9537383B1 (en) | 2015-07-14 | 2015-07-14 | Switching regulator with controllable dead time and quick start |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US14/798,903 US9537383B1 (en) | 2015-07-14 | 2015-07-14 | Switching regulator with controllable dead time and quick start |
Publications (2)
Publication Number | Publication Date |
---|---|
US9537383B1 US9537383B1 (en) | 2017-01-03 |
US20170019015A1 true US20170019015A1 (en) | 2017-01-19 |
Family
ID=57682452
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/798,903 Active 2035-07-30 US9537383B1 (en) | 2015-07-14 | 2015-07-14 | Switching regulator with controllable dead time and quick start |
Country Status (1)
Country | Link |
---|---|
US (1) | US9537383B1 (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180145592A1 (en) * | 2016-08-22 | 2018-05-24 | Ferric Inc. | Zero-Voltage Switch-Mode Power Converter |
US10291128B1 (en) | 2017-12-19 | 2019-05-14 | Linear Technology Holding Llc | Minimizing body diode conduction in synchronous converters |
CN109994999A (en) * | 2017-12-29 | 2019-07-09 | 丰田自动车株式会社 | Protection control device and its control method for power conversion circuit |
US10389257B2 (en) * | 2016-10-21 | 2019-08-20 | Skyworks Solutions, Inc. | Switch efficiency optimization circuit |
US10432092B2 (en) * | 2017-11-17 | 2019-10-01 | Texas Instruments Incorporated | Self-calibrated DC-DC converter |
US10804886B2 (en) * | 2016-06-21 | 2020-10-13 | Eagle Harbor Technologies, Inc. | High voltage pre-pulsing |
US10855270B1 (en) | 2019-10-14 | 2020-12-01 | Silanna Asia Pte Ltd | Adaptive control of non-overlapping drive signals |
DE102022124373A1 (en) | 2022-09-22 | 2024-03-28 | Bayerische Motoren Werke Aktiengesellschaft | Power converter for an electrical machine of a vehicle and method for operating a power converter |
Families Citing this family (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10116211B2 (en) * | 2015-02-11 | 2018-10-30 | Mediatek Inc. | Power converter with adaptive zero-crossing current detection |
US10230356B2 (en) | 2017-02-27 | 2019-03-12 | Allegro Microsystems, Llc | High-side output transistor circuit |
DE102017218305A1 (en) | 2017-10-13 | 2019-04-18 | Conti Temic Microelectronic Gmbh | Method for controlling a semiconductor bridge of an electrically operable motor by means of a ramp signal, control device and arrangement |
US10367500B2 (en) | 2017-12-08 | 2019-07-30 | Allegro Microsystems, Llc | Switching voltage regulator with variable minimum off-time |
EP3719981A1 (en) | 2019-04-04 | 2020-10-07 | Tridonic GmbH & Co. KG | Llc stage for led drivers |
US11695330B2 (en) * | 2019-09-11 | 2023-07-04 | Analog Devices International Unlimited Company | Method to reduce the common-mode EMI of a full bridge converter using sampling common-mode feedback |
US11469669B2 (en) * | 2020-01-31 | 2022-10-11 | Texas Instruments Incorporated | Methods and circuitry to detect PFM mode entry in wide duty range DC converter |
TWI753548B (en) * | 2020-08-26 | 2022-01-21 | 華邦電子股份有限公司 | Low-dropout regulator |
US11711012B2 (en) | 2021-05-28 | 2023-07-25 | Allegro Microsystems, Llc | Synchronous converter for use with reverse current protection diode |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6396250B1 (en) * | 2000-08-31 | 2002-05-28 | Texas Instruments Incorporated | Control method to reduce body diode conduction and reverse recovery losses |
US9425785B1 (en) * | 2015-07-14 | 2016-08-23 | Allegro Microsystems, Llc | Switching regulator with controllable slew rate |
-
2015
- 2015-07-14 US US14/798,903 patent/US9537383B1/en active Active
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10804886B2 (en) * | 2016-06-21 | 2020-10-13 | Eagle Harbor Technologies, Inc. | High voltage pre-pulsing |
US11962304B2 (en) | 2016-06-21 | 2024-04-16 | EHT Ventures LLC | High voltage pre-pulsing |
US11303269B2 (en) * | 2016-06-21 | 2022-04-12 | Eagle Harbor Technologies, Inc. | High voltage pre-pulsing |
US10326366B2 (en) * | 2016-08-22 | 2019-06-18 | Ferric Inc. | Zero-voltage switch-mode power converter |
US20180145592A1 (en) * | 2016-08-22 | 2018-05-24 | Ferric Inc. | Zero-Voltage Switch-Mode Power Converter |
US10389257B2 (en) * | 2016-10-21 | 2019-08-20 | Skyworks Solutions, Inc. | Switch efficiency optimization circuit |
US20200076313A1 (en) * | 2016-10-21 | 2020-03-05 | Skyworks Solutions, Inc. | Switch efficiency optimization circuit |
US10811980B2 (en) * | 2016-10-21 | 2020-10-20 | Skyworks Solutions, Inc. | Switch efficiency optimization circuit |
US10432092B2 (en) * | 2017-11-17 | 2019-10-01 | Texas Instruments Incorporated | Self-calibrated DC-DC converter |
US11356020B2 (en) * | 2017-11-17 | 2022-06-07 | Texas Instruments Incorporated | Self-calibrated DC-DC converter |
US10291128B1 (en) | 2017-12-19 | 2019-05-14 | Linear Technology Holding Llc | Minimizing body diode conduction in synchronous converters |
US10432083B2 (en) * | 2017-12-29 | 2019-10-01 | Toyota Jidosha Kabushiki Kaisha | Protection control apparatus for power conversion circuitry and control method thereof |
CN109994999A (en) * | 2017-12-29 | 2019-07-09 | 丰田自动车株式会社 | Protection control device and its control method for power conversion circuit |
US10855270B1 (en) | 2019-10-14 | 2020-12-01 | Silanna Asia Pte Ltd | Adaptive control of non-overlapping drive signals |
US11611337B2 (en) | 2019-10-14 | 2023-03-21 | Silanna Asia Pte Ltd | Adaptive control of non-overlapping drive signals |
US11855615B2 (en) | 2019-10-14 | 2023-12-26 | Silanna Asia Pte Ltd | Adaptive control of non-overlapping drive signals |
DE102022124373A1 (en) | 2022-09-22 | 2024-03-28 | Bayerische Motoren Werke Aktiengesellschaft | Power converter for an electrical machine of a vehicle and method for operating a power converter |
Also Published As
Publication number | Publication date |
---|---|
US9537383B1 (en) | 2017-01-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9537383B1 (en) | Switching regulator with controllable dead time and quick start | |
US9425785B1 (en) | Switching regulator with controllable slew rate | |
US8436594B2 (en) | Control circuit and method for a digital synchronous switching converter | |
US9595869B2 (en) | Multi-level switching regulator circuits and methods with finite state machine control | |
US9577525B2 (en) | Adaptive dead time control | |
US9621036B2 (en) | Circuits and techniques for improving regulation in a regulator having more than one mode of operation | |
US7960947B2 (en) | Adaptive frequency compensation for DC-to-DC converter | |
US20190267906A1 (en) | Synchronous rectifier off control module and synchronous rectifying control circuit | |
US10199919B2 (en) | Zero dead time control circuit | |
KR101285577B1 (en) | Dc-dc converter controller having optimized load transient response and method thereof | |
US8829872B1 (en) | Systems and methods for dropping and/or adding phases in multiphase regulators | |
US10594315B2 (en) | Switching rate monitoring and control | |
US10622891B2 (en) | Voltage-based auto-correction of switching time | |
US10554126B2 (en) | Continuous comparator with improved calibration | |
US6798178B1 (en) | Method of forming a power system and structure therefor | |
US11392157B2 (en) | Auxiliary apparatus for controlling mode of DC-DC converter | |
US10181786B1 (en) | Anti-cross-conduction time interval minimizer | |
US10164531B2 (en) | Adaptive control method for generating non overlapping time in output devices | |
JP2014161146A (en) | Switching power supply device | |
US10784775B1 (en) | Switching converter with reduced dead-time | |
EP4102720A1 (en) | Actively tracking switching speed control and regulating switching speed of a power transistor during turn-on | |
JP2015006120A (en) | Switching power supply apparatus | |
EP3945678A1 (en) | Actively tracking switching speed control of a power transistor | |
US11271482B2 (en) | DC-DC converter and DC-DC converter operation method | |
US11081957B2 (en) | Power converter with multi-mode timing control |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: ALLEGRO MICROSYSTEMS, LLC, MASSACHUSETTS Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:WIBBEN, JOSHUA;REEL/FRAME:036090/0893 Effective date: 20150713 |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
CC | Certificate of correction | ||
MAFP | Maintenance fee payment |
Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1551); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY Year of fee payment: 4 |
|
AS | Assignment |
Owner name: MIZUHO BANK LTD., AS COLLATERAL AGENT, NEW YORK Free format text: PATENT SECURITY AGREEMENT;ASSIGNOR:ALLEGRO MICROSYSTEMS, LLC;REEL/FRAME:053957/0620 Effective date: 20200930 Owner name: CREDIT SUISSE AG, CAYMAN ISLANDS BRANCH, AS COLLATERAL AGENT, NEW YORK Free format text: PATENT SECURITY AGREEMENT;ASSIGNOR:ALLEGRO MICROSYSTEMS, LLC;REEL/FRAME:053957/0874 Effective date: 20200930 |
|
AS | Assignment |
Owner name: ALLEGRO MICROSYSTEMS, LLC, NEW HAMPSHIRE Free format text: RELEASE OF SECURITY INTEREST IN PATENTS (R/F 053957/0620);ASSIGNOR:MIZUHO BANK, LTD., AS COLLATERAL AGENT;REEL/FRAME:064068/0360 Effective date: 20230621 Owner name: MORGAN STANLEY SENIOR FUNDING, INC., AS THE COLLATERAL AGENT, MARYLAND Free format text: PATENT SECURITY AGREEMENT;ASSIGNOR:ALLEGRO MICROSYSTEMS, LLC;REEL/FRAME:064068/0459 Effective date: 20230621 |
|
AS | Assignment |
Owner name: ALLEGRO MICROSYSTEMS, LLC, NEW HAMPSHIRE Free format text: RELEASE OF SECURITY INTEREST IN PATENTS AT REEL 053957/FRAME 0874;ASSIGNOR:CREDIT SUISSE AG, CAYMAN ISLANDS BRANCH, AS COLLATERAL AGENT;REEL/FRAME:065420/0572 Effective date: 20231031 |
|
MAFP | Maintenance fee payment |
Free format text: PAYMENT OF MAINTENANCE FEE, 8TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1552); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY Year of fee payment: 8 |