You are on page 1of 7

Modeling of V2 Current-Mode Control

Jian Li and Fred C. Lee Center for Power Electronics Systems The Bradley Department of Electrical and Computer Engineering Virginia Polytechnic Institute and State University Blacksburg, VA 24061 USA
Abstract-Recently, the V2 type of constant on-time control has been widely used to improve light-load efficiency. In V2 implementation, the nonlinear PWM modulator is much more complicated than usual, since not only is the inductor current information fed back to the modulator, but the capacitor voltage ripple information is also fed back to the modulator. Generally speaking, there is no sub-harmonic oscillation in constant ontime control. However, the delay due to the capacitor ripple results in sub-harmonic oscillation in V2 constant on-time control. So far, there has been no accurate model to predict instability issue due to the capacitor ripple. This paper presents a new modeling approach for V2 constant on-time control. The power stage, the switches and the PWM modulator are treated as a single entity and modeled based the describing function method. The model for the V2 constant on-time control achieved by the new approach can accurately predict sub-harmonic oscillation. Two solutions are discussed to solve the instability issue. The extension of the model to other types of V2 current-mode control is also shown in the paper. Simulation and experimental results are used to verify the proposed model.1

efficiency, since the switching frequency can be lowered to reduce the switching-related loss, as shown in Fig. 3.

Figure 1. Ripple regulator

I.

INTRODUCTION

Figure 2. V2 control architecture

Current-mode control architecture has been widely used for years [1]-[6]. For practical implementation, the equivalent series resistance (ESR) of the output capacitor can be used as the sensing resistor [7], which means that the output voltage ripple, which includes the current information, can be directly used as the ramp for the duty cycle modulation, as shown in Fig. 1. Later, the control structure can be implemented with an additional compensation loop, which is called the V2 current-mode control [8], [9], as shown in Fig. 2. Based on different modulation schemes, these two architectures consist of constant-frequency peak voltage control, constant-frequency valley voltage control, constant on-time control, constant off-time control and hysteretic control. Among all of these control structures, the constant on-time control is the most widely used to improve light-load

Figure 3. V2 constant on-time control structure

This work was support by Analog Devices, C&D Technologies, CRANE, Delta Electronics, HIPRO Electronics, Infineon, Intel, International Rectifier, Intersil, FSP-Group, Linear Technology, LiteOn Tech, Primarion, NXP, Renesas, National Semiconductor, Richtek, Texas Instruments. This work also made use of Engineer Research Center Shared Facilities supported by the National Science Foundation under NSF Award Number EEC-9731677. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect those of the National Science Foundation.

In the V implementation, the nonlinear PWM modulator becomes much more complicated, because not only is the inductor current information fed back to the modulator, but the capacitor voltage ripple information is fed back to the modulator as well. Generally speaking, there is no subharmonic oscillation in the constant on-time control. However, the delay due to the capacitor ripple results in sub-harmonic oscillation in the V2 constant on-time control, as shown in Fig. 4. The system is stable when used with the OSCON capacitor, because the inductor current information dominates the total output voltage ripple; meanwhile sub-harmonic oscillation occurs when the ceramic capacitor is used with the constant on-time control, since the capacitor voltage ripple is too large. This phenomenon also occurs in peak voltage control [8].

978-1-422-2812-0/09/$25.00 2009 IEEE

298

II. PROPOSED MODELING APPROACH FOR V2 CONSTANT ONTIME CONTROL The V2 constant on-time control is used to exemplify the proposed modeling approach. As shown in Fig. 5, the nonlinear constant on-time modulator consists of switches, the output voltage, the comparator and the on-time generator. Its reasonable to treat these components as a single entity to model instead of breaking them into parts. In the proposed modeling approach, the describing function (DF) method is used to model the non-linear current-mode modulator to obtain the transfer function from the control signal vc to the output voltage vo.

(a)

Modeling of the V2 control becomes a challenge due to the complexity of the PWM modulator. Previously, R. Ridleys model [10] was widely used to design the converter with current-mode control. However, its not reasonable to extend R. Ridleys model to the V2 implementation, since his model is based on discrete-time analysis, which does not consider the influence of the capacitor ripple. This is why the models used in [11]-[13] cant accurately predict the influence from the capacitor ripple in the peak voltage control; all of them are extension of R. Ridleys model. In [14], the KrylovBogoliubov-Mitropolsky (KBM) algorithm is used to improve the accuracy of the model, but numerical computation is required. So far, there is no accurate model for V2 currentmode control. Recently, a new modeling approach has been proposed to model current-mode control [15]. In this modeling approach, the describing function (DF) method [16]-[18] is used to model the non-linear current-mode modulator to obtain the transfer function from the control signal vc to the output voltage vo. A similar methodology can be applied here as well. In Section II, constant on-time control is used as an example to illustrate the modeling process and the prediction of the sub-harmonic oscillation is shown in the proposed model. Then, Section III describes two solutions to solve the instability issue. In Section IV, the extension of this model to other V2 types of current-mode control is presented. Simulation and experiment are used to verify the model in Section V. Finally, a summary is given in Section VI.

(b) Figure 4. Sub-harmonic oscillation in V2 constant on-time control: (a) OSCON capacitor (560F/6m), and (b) ceramic capacitor (100F/1.4m)

Figure 5. Modeling Methodology

As shown in Fig. 5, a sinusoidal perturbation with a small magnitude at the frequency fm is injected through the control signal vc; then based on the perturbed output voltage waveform, the describing function from the control signal vc to the output voltage vo can be found by mathematical derivation. Before applying the DF method, it is necessary to make several assumptions: (i) the magnitude of the inductor current slopes during the on-period and the off-period stay constant separately; (ii) the magnitude of the perturbation signal is very small; and (iii) the perturbation frequency fm and the switching frequency fs are commensurable, which means that Nfs = Mfm, where N and M are positive integers. Following the modulation law of V2 constant on-time control, the duty cycle and the output voltage waveforms are shown in Fig. 6. Because the on-time Ton is fixed, the off-time Toff is modulated by the perturbation signal vc(t): sin(2f m t + ) , where r0 is the steady-state DC vc (t ) = r0 + r value of the control signal, r 0 is the magnitude of the perturbation, and is the initial angle. Based on the modulation law, it is found that:

Figure 6. Perturbed waveform

978-1-422-2812-0/09/$25.00 2009 IEEE

299

vc (t i 1 + Toff ( i 1) ) + s nTon = vc (t i + Toff (i ) ) + s f Toff (i )

ti +Toff ( i )

ti 1 +Toff ( i 1)

[i L (t ) vo (t ) / RL ]dt / C o

(1) where, Toff(i) is the ith cycle off-time, sn = RCo (Vin Vo ) / Ls , s f = RCoVo / Ls , Ls is the inductance of the inductor, RCo is the ESR of the output capacitors, Co is the capacitance of the output capacitors, RL is the load resistor, iL(t) is the inductor current, and vo(t) is the output voltage. Assuming Toff (i ) = Toff + Toff ( i ) , where Toff is the steady state off-time, and Toff(i) is the ith cycle off-time perturbation, ti can be calculated as: ti = (i 1)(Ton + Toff ) + i1 Toff ( k ) . Based on (1), it is found
k =1

that:

s f [(1 +

Toff 2C o RCo

)Ti (1

2Ton + Toff 2C o RCo

) Ti 1 ]
(2)

= [vc (t i 1 + Toff ( i 1) ) vc (t i + Toff ( i ) )] [vc (t i 2 + Toff ( i 2 ) ) vc (t i 1 + Toff ( i 1) )] vc (t i 1 + Toff ( i 1) +

2f m 2 2f m R L C o

) vc (t i + Toff ( i ) +

2f m 2

) +

The Fourier coefficient at the perturbation frequency fm for e j , so the describing function the control signal vc(t) is r from the control signal to the output voltage in the s-domain can be calculated as (8): vo ( s) f s (1 e sTon )(1 e sTsw )[1 + 1/( RLCo s)] = Toff 2Ton + Toff sTsw vc ( s ) s f (8) (1 + ) (1 )e 2Co RCo 2Co RCo Vin RL ( RCoCo s + 1) Ls s ( RL + RCo )Co s + 1 This transfer function is effective at frequencies even beyond half of the switching frequency if there is no outer loop compensation. A similar method in [15] is used to simplify the transfer function as (9): ( RCoCo s + 1) vo ( s ) 1 (9) 2 2 vc ( s ) s s s s (1 + + 2) + ) (1 + Q11 12 Q22 2 where, 1 = /Ton, Q1 = 2/, 2 = Tsw/[( RCoCo - Ton/2) ], and RL>>RCo. The simplification is valid for up to half of the switching frequency. When the duty cycle is relatively small, the transfer function can be further simplified as (10):

) 2f m 2 2f m 2 2f m R L C o The perturbed duty cycle d(t) and the perturbed inductor current iL(t) can be expressed by (3) and (4):

vc (t i 2 + Toff ( i 2 ) +

) vc (t i 1 + Toff ( i 1) +

vo ( s ) s s2 ( RCo Co s + 1) /(1 + + 2) vc ( s ) Q2 2 2

(10)

d (t ) = [u (t ti Toff ( i ) ) u (t t i Toff ( i ) Ton )]


i =1

(3) (4)

iL (t ) = [
0

V Vin d (t ) o ]dt + iL 0 Ls Ls

where, u(t)=1 when t>0, and iL0 is the initial value of the inductor current. Then, the Fourier analysis can be performed on the inductor current:

cm(iL) = j 2 f m / N

t M +Toff ( M ) +Ton

iL (t ) e j 2f mt dt

(5)

where, cm(iL) is the Fourier coefficient at the perturbation frequency fm for the inductor current. Based on the result in [15], the coefficient can be calculated as (6):

cm(iL)

f = s sf

(1 e j 2f mTon )(1 e j 2f mTsw )(1 (1 + Toff 2Co RCo ) (1

2Ton + Toff 2Co RCo

2f m RL Co )e j 2f mTsw

Vin e j Ls j 2f m

(6) where, Tsw is the steady-state switching period. Next, Fourier coefficient cm(vo) of the output voltage vo can be calculated based on (7):

cm ( vo ) = cm ( iL )

RL ( RCo Co j 2f m + 1) ( RL + RCo )Co j 2f m + 1

(7)

From the transfer function, it is clear that the double poles at half of the switching frequency may move to the right halfplane according to different capacitors parameters. The critical condition for stability is RCoCo>Ton/2, which clearly shows the influence of the capacitance ripple. The control-to-output voltage transfer function comparison between the peak current-mode control and the V2 constant-on time is shown in Fig. 7. In the peak current-mode control, the Q factor of the double poles at half of the switching frequency is determined by the duty cycle: if there no external ramp, when the duty cycle is larger than 0.5, two poles will move to the right half-plane and the system becomes unstable. In V2 constant on-time control, the Q factor is not only related to the on-time Ton, but also related to the capacitor parameters. The critical condition RCoCo>Ton/2 reflects the interaction between the ESR and the capacitance of the output capacitor, which means those two parameters must be considered at the same time. Different types of capacitors result in different system performance. When fs = 300 KHz and D 0.1, the parameters of the OSCON capacitors (560F/6m) meet the critical condition, so the system is stable. However, the parameters of the ceramic capacitors (100F/1.4m) cannot satisfy the critical condition, so sub-harmonic oscillation occurs, as shown in Fig. 4. The output impedance can be also derived based on similar methodology. As shown in Fig. 8, a sinusoidal perturbation with a small magnitude at the frequency fm is injected through the output current io, then based on the perturbed output voltage waveform, the describing function from the output current io to the output voltage vo can be found out by mathematical derivation.

978-1-422-2812-0/09/$25.00 2009 IEEE

300

(a) (b) Figure 7. Control-to-output transfer function comparison: (a) peak current-mode control, and (b) V2 constant on-time control

Figure 9. Output impedance

possible solutions are proposed: the first is adding the inductor current ramp; the second solution is adding an external ramp. Detailed analysis of these two approaches is presented below. A. Solution I: Adding the inductor current ramp As discussed above, the capacitor voltage ripple is detrimental to the system stability, so an additional current loop can be introduced to enforce the current feedback information and reduce the influence of the capacitor voltage ripple, as shown in Fig. 10.
Figure 8. Model methodology for output impedance

In the s-domain, the output impedance is derived as (11):


f Z o ( s) = [ s sf (1 e sTon )(1 e sTsw )( RCo + (1 + Toff 2Co RCo ) (1 2Ton + Toff 2C o RCo 1 ) Co s )e sTsw Vin 1 1] ( RCo + ) Ls s Co s

(11) The simplified output impedance is expressed as (12): 1 ( RCo Co s + 1) 1 1] ( RCo + ) Z o (s) [ s s2 s s2 Co s (1 + + ) (1 + + 2) Q11 12 Q22 2 (12) When the duty cycle is relatively small, the output impedance can be further simplified as (13):

Figure 10. Solution I: adding the inductor current ramp

Following the same modeling methodology, the control-tooutput transfer function can be derived as (14):

vo ( s ) vc ( s )

Z o ( s) [

Ton Ton (1 + D 2 ) s ( RCoCo s + 1) (13) ( RCo ) 2 ] s s2 2 2Co 2 Co (1 + + 2) Q22 2

( RCo C o s + 1) s s s s2 (1 + + 2) + 2 ) (1 + Q11 1 Q3 2 2 1
2

(14)

where, Ri is the sensing gain of the additional current loop, s f = ( RCo + Ri )Vo / Ls , and Q3 = Tsw /{[( RCo + Ri )C o Ton / 2] } . By comparing (10) and (14), we see that adding the inductor current ramp equivalently increases the ESR of the output capacitors. The sensing gain of the inductor current Ri can be used as a design parameter to eliminate sub-harmonic oscillation for various output capacitors. The output impedance can also be derived as (15): ( RCoCo s + 1) 1 1 (15) 1] ( RCo + ) Z o (s) [ 2 2 s s s s Co s (1 + + 2 ) (1 + + 2) Q11 1 Q32 2

where, D is the steady-state duty cycle. The Bode plot of the output impedance is shown in Fig. 9. It is found that the output impedance is very low throughout a wide frequency range. This is why this control can deal with the transient response even without the outer loop compensation. III. POSSIBLE SOLUTIONS FOR ELIMINATING THE SUBHARMONIC OSCILLATION In order to eliminate sub-harmonic oscillation due to the capacitor ripple using the V2 constant on-time control, two

978-1-422-2812-0/09/$25.00 2009 IEEE

301

Moreover, when the duty cycle is relatively small, (15) can be further simplified as (16):
(1 + D ) Ton Ton {[ 2 ( Ri RCo )]s + 1} ( RCo C o s + 1) 2 Ri C o 2 Ri 2C o Z o ( s ) Ri s s2 (1 + + 2) Q3 2 2
2

(17):
1 vo ( s) vc (s) (1 + (1 + s s
2

(1 +

s Q1 2 + s

s2

(16) An example using the ceramic capacitors is shown in Fig. 11 and Fig. 12. The parameters are: Vin = 12V, Vo = 1.2V, Ls = 300nH, Ton = 0.33s, and the output capacitor consists of eight ceramic capacitors (100F/1.4m). The control-tooutput transfer function and the output impedance are plotted to show the influence of the inductor current ramp. Comparing Fig. 9 with Fig. 12, it is found that the output impedance stays a constant value Ri at low frequency when there is an additional current loop. In some applications, such as voltage regulators for microprocessors where constant output impedance is required, this control structure can be used to meet the design target. For other applications without such a requirement, the outer loop compensation can be used to lower the output impedance in the frequency range within the control bandwidth. B. Solution II: Adding the external ramp As we know, the external ramp is used to eliminate subharmonic oscillation in the peak current-mode control. A similar concept can be used in V2 constant on-time control, as shown in Fig. 13. The external digital ramp starts to build up at the end of the on-time period and resets at the beginning of the on-time period in every switching cycle.

Q2 2

+ ) Q11 12 s s2 + 2 )(1 +

22
)+

)( RCo C o s + 1)

(17)
s Q1 2
2 2 2

se RCo C oTsw s 2 sf

where, se the magnitude of the external ramp. When the duty cycle and the external ramp are small, (17) can be further simplified as (18):

vo ( s ) vc ( s )

( RCo Co s + 1) s s2 (1 + + 2) Q4 2 2

(18)

where, Q4 = Tsw /{[(2se / s f + 1) RCo Co Ton / 2] } . The Q factor is damped due to the external ramp. The output impedance can be derived as (19):
1 (1 + Z o ( s) [ (1 + s + s
2

(1 +

s Q12

s2
2 2

)( RCoCo s + 1) 1] ( RCo + 1 ) Co s

Q22

Q11 1 s s2 + 2 )(1 +

) 2

s + 2 ) + e RCoCoTsw s 2 Q12 2 sf

s2

(19) Based on the parameters used in the previous section, the control-to-output transfer function and the output impedance are plotted in Fig. 14 and Fig. 15 to show the influence of the external ramp. IV. EXTENSION TO OTHER TYPES OF V2 CURRENT-MODE CONTROL The proposed model strategy can be extend to other types of V2 current-mode control structures, including constant offtime, constant-frequency peak voltage control, and constantfrequency valley voltage control structures. When using constant frequency modulation, an external ramp is added to help stabilize the system. Peak voltage control is used as example for further illustration. The parameters are: input voltage Vin = 12V, switching frequency fs = 300KHz, and external ramp se = 0. The transfer functions for peak voltage control are derived as (20) and (21):

Figure 13. Solution II: adding the external ramp

The control-to-output transfer function can be derived as

Figure 11. The control-to-output transfer function with different Ri

Figure 12. The output impedance with different Ri

Figure 14. The control-to-output transfer function with different external ramps

Figure 15. The output impedance with different external ramps

978-1-422-2812-0/09/$25.00 2009 IEEE

302

v o ( s) v c ( s)

( RCo C o s + 1) ( s f s e ) RCo C o s f Toff / 2 s2 + 2 )(1 + + 2) (1 + Tsw s 2 Q5 2 2 Q1 2 2 sn + s f s s2 s

(20)
Z o ( s) ( RCo [ (1 + s Q5 2 + s
2

1 + ) Co s ( RCo C o s + 1) )(1 + 2 s Q1 2 + s
2

) 2

( s f s e ) RCo C o sn + s f

s f Toff 2 Tsw s 2

1]

voltage control even when D = 0.1. It is clear that subharmonic oscillation is more likely occurs when using peak voltage control due to the influence of the capacitor ripple. Based on the duality principle, the properties of the constant off-time control and the valley voltage control can be easily found based on the previous analysis on the constant on-time control and the peak voltage control. V. SIMULATION AND EXPERIMENTAL VERIFICATION The SIMPLIS simulation tool is used to verify the proposed model for V2 constant on-time control. The parameters of the buck converter are as follows: Vin = 12V, Vo = 1.2V, Ton = 0.33s, fs 300KHz, and Ls = 300nH. The control-to-output transfer function and the output impedance are plotted using the simulation results. As shown in Fig. 18 and Fig. 19, the proposed model can accurately predict the system response. The two approaches to eliminate the sub-harmonic oscillation in V2 constant on-time control are also verified through the SIMPLIS simulation. For the case of adding an inductor current ramp, the control-to-output transfer function and the output impedance are shown in Fig. 20 and Fig. 21. For the case of adding an external ramp, the results are shown in Fig. 22 and Fig. 23. All the results show the validity of the model achieved by the proposed modeling approach.

(21) RCo C o Tsw . where, Q = 1 /[( 1 + ) ] 5 2 Tsw s n /( s n + s f ) + Ton The control-to-output transfer function is plotted based on different duty cycles and different capacitor parameters, as shown in Fig. 16 and Fig. 17. One common characteristic of the peak voltage control and the peak current-mode control is that sub-harmonic oscillation occurs when the duty cycle is larger than a critical value. The difference between these two control structures is that this critical duty cycle value is 0.5 for the peak current-mode control, while this value for the peak voltage control is less than 0.5 which is related to the capacitor parameters. As shown in Fig. 17, different capacitor parameters may result in sub-harmonic oscillation in the peak

Figure 16. The control-to-output transfer function with different duty cycles: capacitor (560F/6m)

Figure 17. The control-to-output transfer function with different capacitor parameters: D = 0.1

Red curve: proposed model; Blue curve: Simplis simulation Figure 18. Control-to-output Figure 19. Output impedance transfer function comparison: comparison: output capacitor output capacitor (56F/6m) (56F/6m)

Red curve: proposed model; Blue curve: Simplis simulation Figure 21. Output Impedance for Solution Figure 20. Control-to-output transfer I (8 output capacitors: 100F/1.4m): Ri function for Solution I (8 output = 1m capacitors: 100F/1.4m): Ri = 1m

Red curve: proposed model; Blue curve: Simplis simulation Figure 22. Control-to-output Figure 23. Output Impedance for transfer function for Solution II (8 Solution II (8 output capacitors: output capacitors: 100F/1.4m) 100F/1.4m)

978-1-422-2812-0/09/$25.00 2009 IEEE

303

Figure 24. Experiment setup (1)

(a) (b) Figure 25. Control-to-output transfer function comparison: (a) RCo = 0.39, Co = 210F, and (b) RCo = 0.39, Co = 20.47F (Dashed line: proposed model; Solid line: measurement)

Figure 26. Experiment setup (2)

(a) (b) Figure 27. Control-to-output transfer function comparison: (a) Ri = 0.15, Co = 210F, and (b) Ri = 0.39, Co = 210F (Dashed line: proposed model; Solid line: measurement) [2] A.D. Schoenfeld and Y. Yu, ASDTIC Control and Standardized Interface Circuits Applied to Buck, Parallel and Buck-Boost DC-to-DC Power Converters, NASA Report NASA CR-121106, February, 1973. [3] C.W. Deisch, Switching Control Method Changes Power Converter into a Current Source, in proc. IEEE PESC78, pp. 300-306. [4] P.L. Hunter, Converter Circuit and Method Having Fast Responding Current Balance and Limiting, U.S. Patent 4,002,963, 1977. [5] L.H. Dixon, Average Current-Mode Control of Switching Power Supplies, Unitrode Power Supply Design Seminar handbook, 1990, pp. 5.1-5.14. [6] R. Redl and N. O. Sokal, Current-mode control, five different types, used with the three basic classes of power converters: small-signal ac and large-signal dc characterization, stability requirements, and implementation of practical circuits, in Proc. IEEE PESC85, pp. 771785. [7] B. P. Schweitzer and A. B. Rosenstein, Free running switching mode power regulator: analysis and design, IEEE Trans. Aerosp., vol. AS-2, pp. 1171-1180, Oct. 1964. [8] D. Goder and W. R. Pelletier, V2 architecture provides ultra-fast transient response in switch mode power supplies," in proc. HFPC96, pp. 19-23. [9] D. Goder, Switching regulators, U.S. Patent, 5,770,940, 1998. [10] R. B. Ridley, A new, continuous-time model for current-mode control, IEEE Trans. Power Electron., vol. 6, no. 2, pp. 271-280, April 1991. [11] W. Huang and J. Clarkin, Analysis and design of multiphase synchronous buck converter with enhanced V2 control, in Proc. HFPC00, 2000, pp. 7481. [12] W. Huang, A new control for multi-phase buck converter with fast transient response, in Proc. IEEE APEC01, Anaheim, California, pp. 273279. [13] S. Qu, Modeling and design considerations of V2 controlled buck regulator, in Proc. IEEE APEC01, Anaheim, California, pp. 507513. [14] J. Sun, Characterization and performance comparison of ripple-based control for voltage regulator modules, IEEE Trans. Power Electron., vol.21, pp. 346 353, March 2006. [15] J. Li and F. C. Lee, New Modeling Approach for Current-Mode Control, in proc. IEEE APEC09. [16] D. M. Mitchell, Pulsewidth modulator phase shift. IEEE Trans. Aerosp., vol. AES-16, pp. 272-278, May 1980. [17] R. D. Middlebrook, Predicting modulator phase lag in PWM converter feedback loops, in proc. Powercon81, paper H-4. [18] J. Sun, Small-signal modeling of variable-frequency pulsewidth modulators, IEEE Trans. Aerosp., vol 38, pp. 1104-108, July 2002.

The experimental verification is done based on the LM34919 evaluation board from National Semiconductor. The experiment setup is shown in Fig. 24, and the parameters are: Vin = 12V, Vo = 5V, Fs = 800kHz, Ls = 15F, Rx = Ry = 2.49K, and RL = 10. Experimental results with different values for Co are shown in Fig. 25. The proposed model can accurately predict the double poles at half of the switching frequency. Experiment verification for adding the inductor current ramp is done based on the same evaluation board with some modifications as shown in Fig. 26. The inductor information is sensed using sensing resistor Ri, and the feedback information is the sum of the inductor current and the output voltage. Experiment results with different values for Ri are shown in Fig. 27. The experiment results agree with the model results very well. VI. CONCLUSIONS This paper presents a new modeling approach for V2 constant on-time control. The power stage, the switches and the PWM modulator are treated as an entity and modeled based the describing function method. The model achieved by the proposed modeling approach can accurately predict subharmonic oscillation due to the capacitor ripple using V2 constant on-time control. Two possible solutions are modeled and analyzed to solve the instability issue. The extension to other types of V2 current-mode control is also demonstrated. Simulation and experimental results are used to verify the proposed model. REFERENCES
[1] L. E. Galllaher, Current Regulator with AC and DC Feedback, U.S. Patent 3,350,628, 1967.

978-1-422-2812-0/09/$25.00 2009 IEEE

304