RTWO-based pulse width modulator

Abstract

A pulse width modulator based on a pair of rotary traveling wave oscillators. The first oscillator operates freely or as part of a phase-locked loop. The second oscillator operates at the same frequency as the first oscillator, but with a controllable phase offset from the first oscillator. The phase offset is set by an input voltage. A block takes the outputs of the first and second oscillators and combines them so that the output is a pulse whose width is the overlap of the oscillation signals from the first and second oscillators. The output pulse width is thus a function of the input voltage. When the pulse width modulator receives the input voltage from the output of a switching power supply, it can use the modulated pulse width to control the switching transistor of the power supply to maintain the output at a regulated voltage.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

1. Field of the Invention

The present invention relates generally to pulse width modulators and more specifically to pulse width modulators employing at least one rotary traveling wave oscillator.

2. Description of the Related Art

DC to DC converters commonly include a switching transistor connected to a filter circuit and a control circuit. FIG. 1A shows a simple converter configuration 100. The converter includes a switching transistor 102, a diode 104, an inductor 106, and a capacitor 108. The switching transistor 102 is connected between the primary energy source 110, such as a battery, and a low pass filter circuit 112, which is formed by the inductor 106 and capacitor 108, as shown. In this type of converter, the inductor 106 also functions as an energy storage element. The capacitor 108 is connected across the load 112 and is used to filter out voltage variations (ripple) at the load 112 that would otherwise occur. Typically, a diode 104 is connected across the filter circuit 113 and the control circuit 114 is connected to the output and to the switching transistor 102.

Such a converter operates with at least two phases, defined by the state of the switching transistor 102. During a first phase when the switching transistor 102 is on, energy is transferred from a primary energy source 110, such as a battery, to a storage element 106, and in this particular case to the capacitor 108 and the load 112. During a second phase when the switch is off, energy is transferred from the storage element 106 to the capacitor 108 and the load 112 and the diode 104 operates to maintain continuity of current.

The control circuit 114 of the converter 100 has two important tasks. It must monitor the voltage vl 118 at the load 112 and it must compare the monitored voltage to a reference voltage. The output of such a comparison is a control voltage vc. Based on the comparison signal vc internal to the control circuit, the control circuit must create a switch control signal d 116 that sets the on and off times during a switching cycle of the switching transistor 102 to compensate for changes in the primary energy source 110 or the load 112. Controlling the on and off times of the switching transistor 102 is called duty cycle control, when the cycle frequency is fixed. A signal derived from the duty cycle control signal d 116 then drives the switching transistor 102. Thus, the control circuit 114 must convert the comparison signal vc to a duty cycle control signal d 116.

A common circuit that performs the conversion from the comparison signal to the duty cycle control signal (vc→d) is a pulse width modulator (PWM). The modulator receives the comparison signal and alters the duty cycle control signal usually in a linear fashion, between a minimum and a maximum duty cycle. For example, if the comparison signal vc ranges from 1.0 to 4.0 volts, the duty cycle control signal d ranges from 10% to 95%. A pulse width modulator operates at some preset switching frequency so that the duty cycle fixes the maximum on and off times of the switching transistor.

Switching frequencies from 100 to 200 KHz are common. Switching frequencies determine the size of the energy storage inductor 106 and filter capacitor 108 shown in FIG. 1A. At 200 KHz, an inductor may have a value of 5 μH and a capacitor a value of 2000 μF. At higher switching frequencies, the values and thus sizes of these components can become much smaller.

However, high switching frequencies affect the power lost in driving the switching transistor. Common types of transistors used for the switching transistor include the MOSFET, the IGBT, and the BJT. When the transistor is a MOSFET, these power losses arise from the input capacitance of the MOSFET. During turn on and turn off, the input capacitance of a MOSFET is a combination of a gate-to-source capacitance Cgs and a gate-to-drain capacitance Cgd, both of which vary with the voltage across them because their capacitance derives in part from the depletion layers in the transistor. The greatest variation in capacitance comes from Cgd, which can vary by a factor of 10 to 100 as a function of the drain-to-gate voltage, V_DG (≈V_DS). Charging and discharging these capacitances causes a power loss in the converter according to the relation CV²f, where V is the voltage output of the driver and f is the switching frequency, and C is the combination of the Cgs and Cgd capacitances at the voltages across them. Values for Cgs might be about 1000 pF and for Cgd about 150 pF to 1500 pF. If the switching frequency is about 200 KHz, then the power lost in charging and discharging the capacitance of a single switching transistor is about 6 to 12 milliwatts (assuming a 5 Volt swing). The losses due to switching rise linearly with the frequency of operation and affect the efficiency of the converter at low load currents.

As mentioned above, a diode is used in combination with the switching transistor to maintain current continuity in the converter. However, diodes also contribute to power losses, thus lowering the efficiency of the converter. For example, if the voltage drop across a diode is about 1 volt with 5 A of current flowing through it, the loss is 5 Watts. This is a serious problem at low loads because the power loss in the diode can be a large fraction of the power delivered to the load, causing low efficiency. Therefore, it is common to replace the diode with a synchronous rectifier. This is shown in FIG. 1B, where a transistor 152 replaces the diode 104 in FIG. 1A and connects via d2156 to the control circuit 154. When the switching transistor 102 is off, the control circuit 154 turns on the synchronous rectifier 152 to maintain continuity of current in the circuit. The synchronous rectifier arrangement substantially lowers the power that would have been dissipated in the diode, because the on-resistance of the transistor 152 can be very small. The on-resistance r_DS(on) of the transistor 152 ranges from 0.01 to 0.1 ohms. However, the additional transistor 152 also causes a frequency-dependent power loss because the control circuit 154 must charge and discharge the input capacitance of the transistor 152.

Also as mentioned above, many converters operate with two phases. Additional phases can improve the converter in at least three ways. First, the additional phases can lower the time to bring the load back into regulation after a significant change in the load occurs. Instead of waiting for multiple switching cycles to occur, the converter need only wait for the additional phases to occur. Second, additional phases can permit the combining of multiple converters so that output currents higher than those obtainable from a single converter are possible. Third, output ripple tends to be smaller in a multiphase converter, making such a converter suitable for a wider variety of applications. However, each additional phase requires at least one additional switch, again increasing CV² f dissipation.

Given the above considerations, it would be desirable to have a very high frequency power converter, synchronously switched and operating with multiple phases.

BRIEF SUMMARY OF THE INVENTION

In one embodiment, the present invention is a pulse-width modulator (PWM) formed from a pair of RTWOs. The RTWOs drive the switches of a power converter.

One advantage of the present invention are that CV²f losses are eliminated. Instead, a small RTWO power overhead is incurred, as the capacitances of the switches become part of the capacitance of the RTWO.

Another advantage is that the RTWOs can operate at very high frequencies, permitting components to become physically much smaller.

Yet another advantage is that synchronous rectifiers are easily driven from the same RTWOs as the switches.

Yet another advantage is that phase shifts that optimally time the synchronous rectifiers are achieved with RTWO phasing and there is no additional CV²f penalty when driven with RTWOs.

Yet another advantage is that poly phase power supplies simply tap uniform phases on the RTWO, with no significant increase in CV²f loss.

Yet another advantage is that devices will operate in series.

BRIEF DESCRIPTION OF THE DRAWINGS

These and other features, aspects and advantages of the present invention will become better understood with regard to the following description, appended claims, and accompanying drawings where:

FIG. 1A shows a conventional step down converter;

FIG. 1B shows a conventional step down converter with a synchronous rectifier;

FIG. 2A shows one embodiment of the present invention;

FIG. 2B shows another embodiment of the present invention;

FIG. 2C shows yet another embodiment of the present invention;

FIG. 2D shows an implementation of the PW block;

FIG. 2E shows an implementation of the PD2, T-V, EA, LPF2, and CTL blocks;

FIG. 3 a conceptual version of the embodiment of FIG. 2C;

FIG. 4A shows a power converter using an embodiment of PWM of the present invention;

FIG. 4B shows a power converter using an embodiment of the PWM of the present invention;

FIG. 4C shows a power converter using yet another embodiment of the PWM of the present invention;

FIG. 4D shows a power converter using yet another embodiment of the PWM of the present invention;

FIG. 4E shows a power converter using yet another embodiment of the PWM of the present invention; and

FIG. 5 shows a multiphase converter.

DETAILED DESCRIPTION OF THE INVENTION

One embodiment 200 of the present invention is a pulse-width modulator. FIG. 2A shows the arrangement of blocks for achieving this function. The arrangement includes a first RTWO1202 and a second RTWO2204. The RTWO1202 may run freely or optionally be phase-locked to a reference such as a crystal oscillator 212. When RTWO1202 operates in a phase-locked loop, blocks PD1206, LPF1208 and CTL1210 are included. Block PD1206 is a phase detector which determines the difference in phase between crystal oscillator output and the RTWO1 output. Block LPF1208 is a low pass filter which averages out the output of the PD block. Block CTL1210 is a control block that adjusts the RTWO1202 so that its frequency and phase matches that of the crystal oscillator 212. The first RTWO 202 sets the operating frequency of the PWM.

The RTWO2204 operates to generate a phase that is offset by a controlled amount from the phase of RTWO1202. The phase offset is controlled by the PD2214, T-V 216, EA 218, LPF2220, and CTL2222 blocks and Vin 224. The PD2 block 214 detects the phase difference between the RTWO1202 and RTWO2204 outputs. The T-V block 216 converts the output of the PD2 block 214 into a voltage. The EA block 218 amplifies (a) the difference (Verror=Vref−Vin) between a reference voltage Vref 226 and the input voltage Vin 224, properly scaled, to create an error voltage Verror, and the (b) difference between the T-V block 216 output and Verror 228. The LPF2 block 220 filters the output of the EA block 218. The CTL2 block 222 alters the phase and/or frequency of the RTWO2204 to have an offset that is, in effect, set by the Verror signal.

FIG. 2E shows one embodiment 250 of the PD2214, T-V 216, EA 218, LPF2220, and CTL blocks 222. The PD2 block 214 includes the exclusive-OR gate (XOR) 252, whose function in the circuit is described in more detail below. The T-V block 216 includes an RC filter 254. The EA block 218 includes a first error amplifier 256 for boosting the difference (Verror) between a scaled Vout and Vref, and a second error amplifier 258 for boosting the difference between the first error amplifier 256 output (Verror) and the filtered output via 254 of the exclusive-OR gate 252. LPF2 block 220 is a conventional low pass filter. In this embodiment, the CTL block 222 is a varactor 260 whose capacitance alters the frequency and phase of the RTWO2204 in FIG. 2A.

The exclusive-OR gate 252 operates to determine the phase difference between the two RTWOs 202, 204. If the RTWOs 202, 204 are in phase alignment (zero phase difference), then the output of the XOR gate 252 is close to zero volts, which is designed to correspond to a no-load condition. If the RTWOs 202, 204 are 90 degrees out of phase, then the output of the XOR gate 252 is approximately VOH/2, which is designed to correspond to a maximum load condition. As described above, the output of the XOR gate 252 is filtered and applied to a first input of the second error amplifier 258, which compares the first input to the second input, received from the first error amplifier 256. Thus, a voltage error (Verror) 262 between the reference 264 and the scaled output 266 creates a finite-size output pulse from the XOR gate 252, which attempts to remove the voltage error 264, via negative feedback.

Block PW 228 in FIGS. 2A and 2B combines selected output taps of RTWO1202 and RTWO2204. The output 229 of the PW block 228 is thus a pulse whose width is modulated by the input Vin 224. FIG. 2D shows an embodiment 270 of the PW block. Instead of using a single switching transistor, two transistors 272, 274 are used in series. The top transistor in the series receives the output of RTWO1202. The bottom transistor receives the output of RTWO2204. The pair of transistors 276 has an on time that is the overlap of the two RTWO outputs, as shown. An alternative to two transistors is a single transistor with dual gates (MOSFET) or bases (BJT), although such transistors are less common.

FIG. 2B shows an alternative design 240 in which the phase offset is digitally controlled. In this embodiment, the operation of RTWO1202 is the same as in the embodiment of FIG. 2A. The blocks for controlling the phase offset derived from RTWO2 are different. These blocks include a PD2 block 242, an LPF2 block 246, a CTL2 block 248, an A/D block 243, and a PS (Phase Selector) block 247. The PD2 block 242 determines the phase difference between RTWO1202 and RTWO2204. The LPF2 block 246 receives the output of the PD2 block 242 and filters it. The CTL2 block 248 adjusts the frequency and/or phase of RTWO2204. The A/D block converts 243 the input voltage to a digital value 245. The PS block 247 selects one of the phases of RTWO2204 according to the digital value 245 from the A/D block 243. Therefore, the Verror signal controls the phase offset from RTWO2204 by selecting it in a phase selector 247. The PW block then combines the output of RTWO1 and RTWO2 to create the desired pulse.

FIG. 2C shows a more simple embodiment 300 for digitally controlling the pulse width, which is illustrated conceptually in FIG. 3. This embodiment includes an A/D block 302, a PS block 304 and a PW block 306, along with a single RTWO 308. The A/D block 302 receives the voltage Verror 310 to generate a digital version [d1, d2, . . . , dn] 312 of the control voltage. The digital version of the control voltage [d1, d2, . . . , dn] 312 operates to select one of N phases tapped directly from the RTWO 308 via the PS block 304. The PW block 306 receives the selected phase 316 from the RTWO 308 along with a phase from a fixed tap 314 of the RTWO and operates to combine the phases to create the desired pulse width.

Power Converter Design

FIG. 4A shows the pulse width modulator used in a power converter 400. The input voltage Vin 402 connects to the output voltage 404 of the converter via a scaler block 406, which may increase or decrease the load voltage 404, and the output of the PW block 408 connects to the duty cycle input d 410 of the converter, as shown.

FIG. 4B shows the pulse width modulator used in a power converter with a synchronous rectifier 104. In FIG. 4B, a second PW block 452 is added to form the duty cycle signal 454 for operating the synchronous transistor 104.

FIG. 4C shows a pulse width modulator used in a power converter 480 with a synchronous rectifier. In FIG. 4C, the output of each RTWO 202, 204 operates the switching transistor 102 and synchronous rectifier 104 directly, eliminating the need for the PW blocks.

FIG. 4D shows a pulse width modulator used in a power converter 500 with a synchronous rectifier 104. The pulse width modulator uses a digital selection 312 for the phase based on the error voltage 310 from the error amplifier 502, in accordance with the design in FIG. 2C. Error amplifier receives input from Vref 504 and the scaler 406 to create the Verror signal 310.

FIG. 4E shows a pulse width modulator used in a power converter 520 with a synchronous rectifier 104. The pulse width modulator uses a digital selection 312 for the phase based on the error voltage 310, in accordance with the design in FIG. 3. In FIG. 4E, it should be noted that not only can the main switching transistor 102 be controlled by the digital phase selector, the synchronous rectifier transistor 104 could also be controlled in the same manner, i.e., by adding a separate phase selector PS circuit 522 that receives the output of the A/D converter 312 and N taps 524 from the RTWO 314.

Multiphase Power Converter

Yet another embodiment 550 of the present invention is a multiphase power converter. A portion of such a converter is shown in FIG. 5. Each of the switches 552a,b 554a,b 556a,b charges and discharges inductors 560, 562, 564 via voltage and current source 564 shown and can be controlled by a PW block derived from pulse width modulator of the present invention or directly from an RTWO, in accordance with FIG. 4A, 4B, 4C, 4D, or 4E. The design 580 in FIG. 3 can also be used as a PWM for a multiphase converter, by using taps 582a-n from the RTWO 584 that are spaced 360/N degrees apart for an N phase converter.

Although the present invention has been described in considerable detail with reference to certain preferred versions thereof, other versions are possible. Therefore, the spirit and scope of the appended claims should not be limited to the description of the preferred versions contained herein.

Claims

1. A pulse width modulator comprising: a first rotary traveling wave oscillator oscillating at a predetermined frequency and having an oscillator signal at a first output tap and a feedback tap with a first phase;a second rotary traveling wave oscillator oscillating at a second frequency and having an oscillator signal at a second output tap and a feedback tap with a second phase;a voltage controlled capacitance device, receiving a voltage control signal that alters the capacitance of said device, said device connected to the second rotary traveling wave oscillator such that the capacitance of said device affects the frequency and phase of said second oscillator;a phase detector that connects to the feedback taps from the first and second rotary traveling wave oscillators to determine any difference between the first phase and the second phase;a time-to-voltage converter that converts any difference from the phase detector into a voltage;an error amplifier that receives the voltage output of the time-to-voltage converter and an external input voltage to generate an error signal;a low pass filter that filters the error signal from the error amplifier to generate the voltage control signal that tends to make the second frequency substantially the same as the predetermined frequency and the second phase adjustably offset from the first phase; anda pulse width device that combines the oscillator signal at the first output tap and the oscillator signal at the second output tap to create a pulse whose width is a function of the input voltage.

2. The pulse width modulator as recited in claim 1, wherein the phase detector is an exclusive ‘OR’ gate having a first input connected to the feedback tap of the first oscillator and a second input connected to the feedback tap of the second oscillator.

3. The pulse width modulator as recited in claim 1, wherein the time-to-voltage converter is a low pass filter that receives the output from the phase detector to create a voltage that averages the phase detector over time.

4. The pulse width modulator as recited in claim 1, wherein the pulse width device includes a first transistor and a second transistor, each having a source, drain, gate, and channel between the drain and source, the channels of said first and second transistors being connected in series, the gate of the first transistor being connected to the first output tap and the gate of the second transistor being connected to the second output tap so that a conductive path through both transistors occurs when the oscillator signal at the first tap and the oscillator signal at the second tap overlap in time.

5. A power converter comprising: a first rotary traveling wave oscillator (RTWO) configured to generate a first oscillator signal;a second RTWO configured to generate a second oscillator signal;a phase detector configured to determine a phase difference between the first and second oscillator signals;a time-to-voltage converter configure to convert the phase difference into a phase voltage;an error amplifier configured to generate an amplified voltage by amplifying a difference between the phase voltage and an error voltage; anda control block configured to control at least one of a phase or a frequency of the second RTWO based on the amplified voltage such that the phase difference between the first and second oscillator signals is controlled by the error voltage.

6. The power converter of claim 5, further comprising a pulse width block configured to generate a pulse signal by combining the first oscillator signal and the second oscillator signal.

7. The power converter of claim 6, wherein the pulse width block comprises a first transistor and a second transistor, wherein the first transistor includes a gate configured to receive the first oscillator signal and a channel, wherein the second transistor includes a gate configured to receive the second oscillator signal and a channel, wherein the channels of the first and second transistors are connected in series so that a conductive path through the first and second transistors is provided when the first oscillator signal and the second oscillator signal overlap in time.

8. The power converter of claim 5, wherein the error voltage corresponds to a difference between a reference voltage and an input voltage.

9. The power converter of claim 8, further comprising at least one switching transistor configured to control an output voltage of the power converter, wherein the power converter further comprises a scaler block configured to generate the input voltage by scaling the output voltage of the power converter.

10. The power converter of claim 9, wherein the first oscillator signal and the second oscillator signal operate to control a duty cycle of the at least one switching transistor.

11. The power converter of claim 5, wherein the first RTWO is configured to operate in a phase-locked loop such that the first oscillator signal is phase-locked to a reference oscillator signal.

12. The power converter of claim 5, wherein the phase detector comprises an exclusive-OR (XOR) gate having a first input configured to receive the first oscillator signal and a second input configured to receive the second oscillator signal.

13. The power converter of claim 12, wherein the time-to-digital converter comprises a resistor-capacitor (RC) filter configured to generate the phase voltage by filtering an output of the XOR gate.

14. The power converter of claim 5, wherein the control block comprises a varactor having a capacitance configured to alter the phase and frequency of the second RTWO.

15. The power converter of claim 5, further comprising a low pass filter configured to filter the amplified voltage.

16. A method of controlling a duty cycle in a power converter, the method comprising: generating a first oscillator signal using a first rotary traveling wave oscillator (RTWO);generating a second oscillator signal using a second RTWO;determining a phase difference between the first and second oscillator signals;converting the phase difference into a phase voltage;generating an amplified voltage by amplifying a difference between the phase voltage and an error voltage; andcontrolling at least one of a phase or a frequency of the second RTWO based on the amplified voltage.

17. The method of claim 16, further comprising generating a pulse signal by combining the first oscillator signal and the second oscillator signal.

18. The method of claim 16, further comprising controlling a duty cycle of a switching transistor using the pulse signal.

19. The power converter of claim 16, wherein the error voltage corresponds to a difference between a reference voltage and an input voltage.

20. The power converter of claim 19, further comprising scaling an output voltage of the power converter to generate the input voltage.