You are on page 1of 5

This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TCSII.2021.3080289, IEEE
Transactions on Circuits and Systems II: Express Briefs

An Improved Voltage-Mode Controller for the


Quadratic Boost Converter
Chok-You Chan*
Abstract- An improved voltage-mode controller (VMC) for the uses the output voltage for control purpose. It was then used
single-switch conventional quadratic boost converter (CQBC) is in the control of a high-order boost converter and a CQBC,
presented. It has a structure different from those of the existing with the incorporation of proportional and integral actions [19,
voltage-mode controllers in that it avoids the possibility of 15]. Later, the voltage-mode controller was modified to avoid
dividing by zero and finding the square root of a term which
could be negative. It also now uses a proportional term with a
the possibility of dividing by zero in the regulation of the re-
normalized output voltage error as opposed to just using the lift circuit of super-lift converters [16] and the CQBC [17].
normalized output voltage error in the integral term to provide Recently, there has been some effort focused on improving
integral action. Using these normalized signals helps to reduce the the performance of the existing controllers for the high-order
variations in the control signal, leading to improved performance. boost converters, including the CQBC [7, 13-14, 16-17]. The
A comparative experimental study of the proposed VMC and the key feature in these controllers is the use of the normalized
state-of-the-art modified current-mode controller is carried out output voltage error in the integral term to reduce variations in
under various operating conditions. The results are presented in the control signal when there are changes in the load, the input
the brief. and reference voltages, leading to improved performance.
Motivated by these works, this brief focuses on the problem
I. INTRODUCTION of regulating the more well-known and widely studied CQBC
In the past few decades, many applications that require using a VCM when the load resistance is uncertain. In this
power supplies with high conversion ratios have emerged, brief, some modifications to the structure of the existing
particularly in renewable energy sources. The output voltages controllers of [15, 17] to avoid division by zero and finding the
of energy resources like batteries, fuel cells and photovoltaic square root of negative values in the control laws as well as to
cells are rather small and require a dc-dc boost converter with achieve improved performance are addressed. Specifically,
sufficient voltage gain before interfacing with the grid. This unlike the cases in [15, 17] where the normalized output
quickly led to the development of many non-traditional dc-dc voltage error is only used in the integral term to provide
converters with transformers and transformer-less structures integral action, the proportional term now also uses a
which can provide high dc-gains at low values of duty ratio. normalized output voltage error. The use of these normalized
These converters which can provide higher conversion ratios signals helps to reduce the variations in the control signal
are known as high gain boost converters. A comprehensive further and leads to improve performance. A stability analysis
review of voltage-boosting techniques and topologies on dc- is carried out to gain some understanding into the performance
dc converters is given in [1]. Some recent topologies are given of the resulting closed-loop system. The comparative
in [2] – [4]. experimental study of the proposed VMC and the state-of-the-
In the last few years, the regulation of these high-order boost art modified current-mode controller [7] was carried out under
converters, especially the CQBCs, has attracted quite a lot of different operating conditions.
attention due to their more complex structures as compared to
II. MODEL OF THE CQBC
that of the traditional boost converter, the non-minimum phase
characteristic, and the need to use low-order controllers for
Fig. 1 shows the circuit of the CQBC. If operating in the
their easy implementations. Thus far, there are several
continuous mode, its averaged model can be described by [5]:
reported solutions to the problem of regulating these high-
order boost converters. The approaches reported to name a
few, are current-mode control, adaptive PI control, hysteresis- =− (1 − ) + (1)
based and PWM-based sliding-mode controls, voltage-mode
control [5-17]. Most of these approaches use inductor =− (1 − ) + (2)
current(s) and the output voltage for regulation purposes while
the voltage-mode approach only uses the output voltage for the
same purpose. = (1 − ) − (3)
It appears that there has not been much work reported on the
voltage-mode control of high-order boost converters. A
simplified parallel-damped passivity-based controller for the = (1 − ) − (4)
conventional boost converter was first reported in [18]. It only
where is the average current of inductor , is the
*CY Chan is with the School of Electrical & Electronics Engineering,
Nanyang Technological University, Singapore (E-mail: ecychan@ntu.edu.sg)
average current of inductor , is the average voltage of

1549-7747 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Universidade Tecnologica Federal do Parana. Downloaded on June 01,2021 at 17:04:09 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TCSII.2021.3080289, IEEE
Transactions on Circuits and Systems II: Express Briefs

capacitor , is the average voltage of capacitor , and u in the integral action as in [15, 17]. It is worth pointing out
is the duty ratio of the switch S. that the proposed improved VMC described by (6) – (8) is
quite different from those of [15, 17]. In this case, the signals
D1 are smaller than those without normalization when the output
voltage error is large in the presence of circuit parameter
variations and will give improved performance.
L1 L2
D2 vc1 D3 vc2
B. Stability Analysis
iL1 iL2

E C1 S C2 R To carry out the stability analysis of the proposed voltage-


mode controlled system, the following errors are defined:

= − , = − , = − ,
Fig. 1. Single-switch conventional quadratic boost converter.
= − , = − , (9)
Using (1) – (4), the equilibrium values of = , = ,
= , = = and = can be obtained as: Also, define

√ ∫
= , = , = , =1− (5) =1− =1− (10)

where is the desired value of the output voltage . Using (6) – (10) in (1) – (4) yields the following error
dynamics:
III. PROPOSED VMC FOR THE CQBC
=− ( + )+ (11)
This section presents the proposed improved VMC for the
CQBC and the stability analysis of the voltage-mode =− ( + )+ ( + ) (12)
controlled system as well as the feasibility of the VCM.

A. Proposed Controller Structure = ( + )− ( + ) (13)

The existing VMCs for the high-order boost converters use


the basic structure of the VMC for the traditional boost = ( + )− ( + ) (14)
converter [18], i.e., the controller is chosen based on the form
of the equilibrium value of the duty ratio function of the = − .2 (15)
converter in question. The equilibrium of the duty ratio
function of the CQBC is given in (5). A judicious choice for = , where =∫ (16)
with the incorporation of both proportional and integral
actions would be: The desired equilibrium point of (11) – (16) is given by:

= [√ − √ − − ∫ ] (6) ( , , , , , ) = (0, 0, 0, 0, 0, 0) (17)

= [−2 + + ] (7) As proving the stability of the error dynamics (11) - (16) is
rather difficult, a linearization approach is adopted.
Linearizing (11) – (16) around the equilibrium point (17)
where yields the following linear system [20]:
= − , = (8)
= (18)

and , and are the controller gains and > 0 is a


where = [ ] , = − , =
constant. It is worth pointing out that division of zero like in
1, … ,5, = − and
[15] and finding the square root of the term ( + +
∫ ) in the existing voltage-mode controllers [15, 17]
are avoided. If the latter term is negative, the generation of the
control signal might be an issue. In this brief, a normalized
signal in the proportional term, i.e., is used and not just

1549-7747 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Universidade Tecnologica Federal do Parana. Downloaded on June 01,2021 at 17:04:09 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TCSII.2021.3080289, IEEE
Transactions on Circuits and Systems II: Express Briefs

0 0 The movement of the closed-loop poles starting at =


⎡ 0 0 ⎤
0.0065 is indicated by the arrows. The system is stable as all
⎢ ⎥
0 the poles lie in the open left-half complex plane.
=⎢ ⎥
⎢ 0 0 ⎥
⎢ 0 0 0 0 ⎥ C. Feasibility of the VMC
⎣ 0 0 0 1 0 0 ⎦
As the structure of the proposed VMC is different from the
existing ones, its feasibility needs to be investigated. Using (6),
=− , =− √ , = ,
can be written as

=− √ , = , =− ( + ), =( +√ + + ) (21)

= , =− , = , =− , Using (6) - (8), (16) and (21) in the time derivative of gives:
/ /
= , =− , = , = , = { − − }

= { x
= ( − 1) , =− , = , √

[−2 ( +√ + + ) + +
= , =− .
2 ]− − } (22)
Now, the eigenvalues of the matrix in (18) are given by: ( ) .

( )=| − |= + + ⋯+ + =0 (19) where from (14),

where is a complex variable and the coefficients are given = . {1 − [ + −√ − −


in terms of the circuit parameters and controller gains. Due to
space constraint, their expressions, are not presented. − ] }( + )− ( + ) (23)
To ensure the stability of the linear system (18), the values
of the controller gains , and must be chosen such that Setting , , and to their equilibrium values gives the
the eigenvalues of the matrix , i.e., the roots of ( ) all lie following internal dynamics of the closed-loop system:
in the open left-half complex plane. Since the order of ( ) is
high, the root locus method [21] is used to determine the range = { x[− +√ ) + –
of a particular controller gain for fixed values of the other √

controller gains such that the system (18) is stable.


For illustration, consider these circuit parameter values. [ (1 − ) − ]} (24)

= 15 , = 103 , = 470 , = 3.3 ,


The condition for = 0 is that the terms in [·] are equated
= = 100 , =1 Ω (20)
to zero to give the feasible equilibrium point = 1− .
The root locus plot of ( ) for = 0.04, = 0.1 and
For illustration, consider the circuit parameter values in
0.0065 ≤ ≤ 0.06 is shown in Fig. 2.
(20) with = 0.03. Fig. 3 shows the phase-plane diagram
of (24), where is the unique stable equilibrium point.

Fig. 2. Root locus plot for 0.0065 ≤ ≤ 0.06. Fig. 3. Phase-plane diagram of (24).

1549-7747 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Universidade Tecnologica Federal do Parana. Downloaded on June 01,2021 at 17:04:09 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TCSII.2021.3080289, IEEE
Transactions on Circuits and Systems II: Express Briefs

The main result is now given in the following proposition. to give the “most satisfactory” responses. It is seen that the
performance is satisfactory. The output voltage was almost
Proposition: For a desired output voltage > > 0, the able to track the increasing with no overshoots, although
controller (6) – (8) with appropriate values of , and there was an undershoot for the decreasing . Also, the output
locally asymptotically stabilizes the converter (1) – (4) was quickly restored to the desired value after the onset of the
towards the desired equilibrium point ̅ , ̅ , , = parameter changes, with very small deviations.

( , , , ) for any 0 < < ∞. B. Modified Current-Mode Controller (MCMC)


The MCMC implemented is described by [7]:
IV. EXPERIMENTAL RESULTS

In this section, some experimental results showing the =1− − − − ∫


features of the proposed VMC are presented. A prototype of
the CQBC was built using the values of the circuit parameters − (27)
given in (20) and operated at a switching frequency of 30 kHz.
The proposed controller implemented is described by: where , and are fixed nominal values of ,
and = , respectively. The values of the controller gains
= [ − − − ∫ ] (25) used were: = 1.716, = 35.5, = 2.8, = 1.2.
Fig. 6 shows the system responses under the same operating
conditions experienced by the VMC. The performance of the
= [−2 + +2 ] (26) VMC is comparable to that of the MCMC, except that there
are some discernible differences. They are shown in Table 1.
where = , = , = and 0 < < 1 is the
scaling factor. For the experiments, = 0.052 was used. TABLE 1
Fig. 4 shows the block diagram of the implementation of the COMPARISON OF VCM AND MCMC
proposed VMC. The controller was built using only analog
devices like Op Amps and AD633 chips. Mathematical Settling time (ms) for changes in , , VCM MCMC
functions like multiplication, division, square and square roots Δ : 0 –103 V 80.196 68.16
were all implemented using the AD633 chips. Δ : 103 – 72 V 103.12 101.09
Δ : 1000 – 595 Ω 25.96 163
+ 4 4
Squ
( 4 )2
4
Δ : 595 – 1000 Ω 25.435 200
- - + + Div
+
1+( 4 )2
Δ : 15 – 13.3 V 27.6 15.43
1

The MCMC has shorter settling times to reach and stay at


for Δ and Δ . For the VMC, an undershoot exists when
+
2 + 1 ∫ - = [·]/ was changed from 103 V to 72 V and none for the MCMC.

- 2 - Div However, for Δ , the VCM is superior to the MCMC, as it is
+ [·]
2 √∙ - independent of the load resistance (see (6) – (7)). Note that the
√∙ VCM uses only the output voltage for control purposes.
√∙ However, there exists a tradeoff in the MCMC, namely, to get
a good transient response for Δ , the convergence of the
output voltage to will be affected for Δ or vice versa.
Fig. 4. Block diagram of controller implementation.
V. CONCLUSION
An experimental comparative study involving the proposed
VMC and the modified current-mode controller was also An improved VCM was proposed for the CQBC which
carried out. The operating conditions were as follows: (a) avoids division by zero and finding the square root of a term
change in from 0 to 103 V and then 103 to 72 V; (b) change in the existing controllers. It also uses a proportional term with
in from 1000 to 595 Ω and then to 1000 Ω; (c) change in a normalized signal in the control law which leads to improved
from 15 V to 13.3 V and to 12.5 V. performance. The work represents an improvement over the
existing VCMs for the CQBCs and other high-order boost
A. Proposed Voltage-Mode Controller converters. The effectiveness of the VCM was demonstrated
Fig. 5 shows the system responses under various operating experimentally under different operating conditions and its
conditions. The controller gains used were: = 0.226 , performance is comparable to that of the MCMC, even though
= 39.3, = 0.0028 and = 0.12. They were chosen only using the output voltage for control purposes. Its structure
could also be extended to other higher-order boost converters.

1549-7747 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Universidade Tecnologica Federal do Parana. Downloaded on June 01,2021 at 17:04:09 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TCSII.2021.3080289, IEEE
Transactions on Circuits and Systems II: Express Briefs

ACKNOWLEDGEMENT REFERENCES

The author would like to acknowledge the contribution of [1] M. Forouzesh, et al., “Step-Up DC–DC Converters: A Comprehensive
Review of Voltage-Boosting Techniques, Topologies, and
Mr. Lim Yuan Jun for building the converter systems and
Applications,” IEEE Trans. Power Electron., vol. 32, no. 12, pp. 9143
carrying out the experiments. – 9178, Dec. 2017.
[2] M. R. Banaei and H. A. F. Bonab, “High efficiency nonisolated buck-
boost converter based on ZETA converter,” IEEE Trans. Ind. Electron.,
vol. 67, no. 3, pp. 1991-1998, Mar. 2020.
[3] B. Zhu, H. Wang, and D.M. Vilathgamuwa, “Single-switch high step-
up boost converter based on a novel voltage multiplier,” IET Power
Electron., vol. 12, no. 14, pp. 3732-3738, Nov. 2019.
[4] M.R. Banaei and S.G. Sani, “Analysis and implementation of a new
sepic-based single-switch buck-boost dc-dc converter with continuous
input current,” IEEE Trans. Power Electron., vol. 33, no. 12, pp. 10317-
10325, Dec. 2018.
[5] J. Leyva-Ramos, et al., “Switching regulator using a quadratic boost
(upper): 15 V/div, 200 ms/div : 15 V/div, 200 ms/div converter for wide DC conversion ratios,” IET Power Electron., vol. 2,
(lower): 500 mV/div, 200 ms/div no. 5, pp. 605– 613, Sep. 2009.
(a) (b) [6] J. A. Morales-Saldana et al., “Multiloop controller design for a quadratic
boost converter,” IET Elect. Power Appl., vol. 1, no. 3, pp. 362–367,
May 2007.
[7] C. Y. Chan, S. Chincholkar, and W. Jiang, “A modified fixed current-
mode controller for improved performance in quadratic boost
converters,” IEEE Trans. Circuits Syst. II, Exp. Briefs, vol. 67, no. 10,
pp. 2014 – 2018, Oct. 2020.
[8] M. Hernandez-Gomez et al., “Robust Adaptive PI stabilization of a
quadratic converter: Experimental results,” in Proc. IEEE Int. Sym. Ind.
Electron., 2010, pp. 2999-3004,
[9] Y. Jiao, F.L. Luo, and M. Zhu, “Generalised modelling and sliding mode
control for n-cell cascade super-lift DC – DC converters,” IET Power
: 15 V/div, 200 ms/div , : 200 mA/div, 20 µs/div Electron., vol. 4, no. 5, pp. 532-540, May 2010.
(upper trace), (lower trace) [10] J. A. Morales-Saldana et al., “Modelling and control of a DC-DC
(c) (d) quadratic boost converter with R2P2,” IET Power Electron., vol. 7, no.
1, pp. 11–22, Jan. 2014.
[11] O. Lopez-Santos et al., “Robust sliding-mode control design for a
Fig. 5. System responses obtained using the VMC. (a) change in ; voltage regulated quadratic boost converter,” IEEE Trans. Power
(b) change in ; (c) change in ; (d) , waveforms for = 595 Electron., vol. 30, no. 4, pp. 2313- 2327, Apr. 2015.
Ω and = 103 V. [12] S. H. Chincholkar and C. Y. Chan, “Design of fixed-frequency pulse-
width modulation-based sliding-mode controllers for the cascade boost
converter,” IEEE Trans. Circuits Syst. II, Exp. Briefs, vol. 64, no. 1, pp.
51 – 55, Jan. 2017.
[13] S. Chincholkar, W. Jiang, C. Y. Chan, “An improved PWM-based
sliding-mode controller for a DC-DC cascade boost converter,” IEEE
Trans. Circuits Syst. II, Exp. Briefs, vol. 65, no. 11, pp. 1639-1643,
Nov. 2018.
[14] S. Chincholkar, W. Jiang, C. Y. Chan, “An improved PWM-based
sliding-mode controller for a DC-DC cascade boost converter,” IEEE
Trans. Circuits Syst. II, Exp. Briefs, vol. 65, no. 1, pp. 1639-1643, Nov.
2018.
(upper): 15 V/div, 200 ms/div : 15 V/div, 200 ms/div [15] C. Y. Chan, “Investigation of a voltage-mode controller for a cascade
(lower): 500 mV/div, 200 ms/div boost converter,” IET Power Electron., vol. 7, no. 8, pp. 2060-2068,
(a) (b) Aug. 2014.
[16] W. Jiang, S. Chincholkar, C. Y. Chan, “Improved output feedback
controller design for the super-lift re-lift Luo converter,” IET Power
Electron., vol. 10, no. 10, pp. 1147-1155, Jun. 2017.
[17] W. Jiang, S. Chincholkar, C. Y. Chan, “A modified voltage-mode
controller for the quadratic boost dc-dc converter with improved
performance,” IET Power Electron., vol. 11, no. 14, pp. 2222-2231,
Oct. 2018.
[18] C. Y. Chan, “Simplified parallel-damped passivity-based controllers for
dc-dc converters,” Automatica, vol. 44, no. 11, pp. 2977-2980, Nov.
2008.
: 15 V/div, 200 ms/div , : 200 mA/div, 20 µs/div [19] C. Y. Chan, “Analysis and experimental study of an output feedback
(upper trace), (lower trace) controller for a high-order boost dc–dc converter.” IET Power
(c) (d) Electron., vol. 6, no. 7, 1279-1287, Aug. 2013.
[20] H. Khalil, Nonlinear Systems, 3rd ed. Upper Saddle River, NJ, USA:
Pearson Educational International, 2000.
Fig. 6. System responses obtained using the MCMC. (a) change in [21] K. Ogata, Modern Control Engineering, 3rd ed. Upper Saddle River, NJ,
; (b) change in ; (c) change in ; (d) , waveforms for = USA: Prentice-Hall, 1997.
595 Ω and = 103 V.

1549-7747 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Authorized licensed use limited to: Universidade Tecnologica Federal do Parana. Downloaded on June 01,2021 at 17:04:09 UTC from IEEE Xplore. Restrictions apply.

You might also like