I.

INTRODUCTION

Solar cells are known to be nonlinear, and there

exists one operating point (V,,Z,> where the solar
Design and Analysis of a array produces maximum power, as illustrated in
Fig. 1. This nonlinear characteristic has resulted in
Regulated Peak-Power Tracking a number of power system configurations [ 1, 3, 51
that make best use of the solar array. Existing peak-
System power tracking (PPT) systems such as the series
and the parallel are shown in Fig. 2. Since the
battery voltage depends on its depth of charge and
the solar array peak-power voltage varies with
temperature level, illumination level, and age of the
PHUONG T. HUYNH, Member, IEEE
Cadence Design Systems solar array, one has to cascade a power converter
(series regulator) at the system output to obtain bus
BO H. CHO, Senior Member, IEEE
Seoul National University regulation. Simple analysis of the solar array power
processed by all the intermediate power converters
of both regulated series and parallel PPT systems
shows that these two systems do not yield optimum
A regulated peak-power tracking (RF'PT) system for space efficiency. Instead, by combining the series and the
power application is proposed. Large-signal stability analysis
parallel features, an optimal system is formed, as
shown in Fig. 3. Let 7 represents the normalized
is provided to understand the main four different modes of
operations of the system, as well as the mode transitions. A simple
voltage
and effective control scheme for the system is also proposed.
current, current source __source
-A
Small-signal analysis is performed thereafter to provide design power

optimization, and the predictions are verified by computer

simulations.

t
voltage

Fig. 1. Solar array output characteristics.

f5jiy'I

--
Regulator
Regulator

Peak-Power
Tracker
Solar

Series-Configuration Parallel-Configuration

Fig. 2. Existing PPT systems.

bus
Series
Regulator

Manuscript received March 1, 1996; revised August 22, 1996 and

r;tl I*
Array
&J
January 12, 1998.
IEEE Log NO. T-AES/35/1/01491. Fig. 3. Series/parallel PPT system
Authors' addresses: P. T. Huynh, Cadence Design Systems, 6760
Alexander Bell Dr., Suite 250, Columbia, MD 21046; B. H. Cho, efficiency of any power converter (7 < l), and r
Dept. of Electrical Engineering, Seoul National University, 56-1 denotes the ratio of eclipse to orbital day ( r 5 l), a
Shinrim-dong, Kwanak-Ku, Seoul, Korea. crude approximation of the efficiencies of the three
systems is shown in Table I. For the series PPT
0018-9251/99/$10.00 @ 1999 IEEE system, since the solar power to the load is always

84 IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS VOL. 35, NO. 1 JANUARY 1999
 TABLE I TABLE I1
Efficiency Comparison Functions of Component Blocks

Series Parallel SeriesParallel Mode of Series Battery Battery

Operations Regulator Charger Discharger
Efficiency T~ ~ ( 1 Y)
- + rq3 ~ ( 1 Y)
- + rv2
ECLIPSE Off Off Regulate bus
PPT discharge PPT Off Regulate bus
PPT charge Regulate bus PPT Off
Dischurger Trickle charge Regulate bus Trickle charge Off

Fig. 4. Eclipse mode.

Reguletor
Battery Voltage

processed by two converlers, the efficiency is q2. For

the parallel configuration, the solar array sources the
load via the series regulator and charges the battery
Fig. 5 . PFT discharge mode.
via the peak-power tracker during the orbital day
(1 - Y), therefore, the efficiency during day light is
v( 1 - r ) . However, during eclipse, the solar energy
stored in the battery must be reprocessed through
two converters to reach the load; thus, the efficiency Voltage
Load
loop
during eclipse is v3r. The series/parallel PPT system
(RPPT) has better efficiency in the overall because
the solar array power is processed through only one
Fig. 6. PPT charge mode.
converter (series regulator) during the orbital day,
and two converters (battery charger and discharger)
during eclipse. For low E,arth orbits (LEO), the Series
Regulator
efficiency of the RPPT is always better than those of
Voltage
the series and the parallel PPT systems; whereas for Battery
Solar Charger
higher orbits, such as geostationary orbits, the eclipse
period shortens (Y + 0), and the RPPT has the same
performance as the parallel configuration. Thus, the
seriedparallel configuration is applicable to LEOS Fig. 7. TC mode
where use of the solar energy can be maximized.
Furthermore, this system configuration is attractive disabled, as shown in Fig. 6. More importantly, to
because it provides regulated bus while allowing maintain bus regulation, the series regulator switches
modularity and multiple battery configuration, and from solar-output regulation to bus regulation. Finally,
does not involve complicated control schemes. when the battery is fully charged, the charger trickle
There exist three different building blocks that charges the battery to compensate for the battery
govern the operations of the system: the series leakage (trickle-charging (TC) mode), as shown in
regulator, the battery discharger, and the battery Fig. 7. Notice that use of the battery charger and
charger. During a complete orbital revolution, the discharger is complementary, implying that they both
RPPT system can operate on four distinct modes. can be combined into a single hardware block; thereby
Different functions of the series regulator, the battery reducing weight and hardware cost.
discharger, and the battery charger for each mode Various functions that each building block
are summarized in Table 11. During eclipse, the must perform during each mode seem to enhance
battery discharger regulates the bus by sourcing the the complexity of the system. The first foreseen
power from the battery, as illustrated in Fig. 4. As the complexity lies in the control structure of the system.
spacecraft enters sunlight, the solar array is forced to Furthermore, emphasis must also be stressed on
operate at the peak-power point by the series regulator the mode transitions and the control issue thereby.
(Fig. 5 ) since its output power is less than the load Nonlinearities in the solar array and the load can
demand, while the battery discharger continues to result in multiple equilibrium points, and if system
regulate the bus by sourcing the additional required control is not implemented properly, the system can
power. As the illumination level increases and the switch to an undesired operating point [2, 31. First,
solar array maximum power exceeds the load demand, large-signal stability needs to be ensured. Then a
the solar array can source the load and charge the particular RPPT system configuration is proposed
exhausted battery. The battery charger is then activated along with a control scheme. Small-signal analysis
and peak-power tracks, and the battery discharger is is performed thereafter to provide optimum design.

HUYNH & CHO: DESIGN AND ANALYSIS OF A REGULATED PEAK-POWER TRACKING SYSTEM 85
 current source 112 line
source
line 2
Solar Series
Array Regulator
Battery Battery
source load line
Discharger current
line 1

T L
Vbus voltage
Vbat voltage
Fig. 8. System decoupling in PPT discharge mode.
Fig 10 System decoupling in TC mode

source line load line current

extracting excess power to charge the battery, thus
JL source
providing system stability.
charge
Series
Regulator
When the battery charge current is regulated at
load line
constant current during the TC mode, the load line of
Battery
the battery charger seen by the rest of the system is
a constant power load line. It can be combined with
the series regulator load line to form a single constant
Vpeak voltage power load line as illustrated in Fig. 10(b). As a
Fig. 9. System decoupling in PPT charge
result, one obtains three equilibrium points among
which B has been shown to be an unstable operating
point 121. The only desirable equilibrium point is C
II. LARGE-SIGNAL STABILITY ANALYSIS at which both battery charger and the series regulator
can maintain closed-loop control; whereas the system
Nonlinearities in the solar array output is not allowed to operate at point A. It is shown that
characteristic and different building blocks of the under mode transition, the system operating point will
system often result in multiple equilibrium points, move to the desirable operating point C.
among which only one is desirable. The need to
ensure system stability in the large-signal sense A. Transitions and Mode Changes
requires some basic understanding of large-signal
behavior of the system. The mode transitions can be described as
During the eclipse mode, as in Fig. 4, the battery follows. Whenever there is not enough power for
discharger behaves as a stiff voltage source, therefore, the solar array to source the load, the series regulator
maintaining system stability. When the series regulator peak-power tracks and the battery discharger regulates
starts to peak-power track as shown in Fig. 8(a), the bus (PPT discharge mode). When the capacity of
the solar array operates at its peak-power point; and the solar array exceeds the load demand, a negative
therefore, source line 1 must be a constant power current is forced into the battery discharger and
source line. Source line 2, a stiff voltage source, automatically turns it off. By then, the bus voltage
represents the output of the discharger. Assuming a increases because of the excess power from the solar
constant power load line, it is realized that without the array. Consequently, the series regulator switches
battery discharger, the system will become inoperable. to the bus regulation mode and the battery charger
The battery discharger regulates the bus by providing is automatically turned on to peak-power-track and
an appropriate amount of current to compensate for provide rapid charge to the battery. It follows that
the discrepancy between the source line 1 and the the system operation is similar to that from the
load line, as illustrated in Fig. 8(b). Thus only one series-configuration PPT system [ 13. The two main
equilibrium point exists. modes are PPT and TC modes. The PPT mode is
During the PPT charge mode, the battery charger extended into PPT charge and PPT discharge. The
peak-power tracks, as shown in Fig. 9(a). The source transition from PPT discharge to PPT charge or vice
line is therefore a combination of the stiff voltage sink versa doesn't cause any stability problem, since the
and the solar array output characteristic, as depicted peak-power tracker can always bring the operating
in Fig. 9(b). Meanwhile, since the series regulator point to the peak-power point.
maintains bus regulation, the load line is a constant As the battery depth of charge reaches its
power load within operable input voltage range and maximum level, the charger trickle-charges the battery
becomes resistive under lower input voltage range at minimum constant current to compensate for the
as the series regulator loses its closed-loop control. battery leakage current. Thus, the solar array output
Consequently, three different equilibrium points result, voltage is neither regulated by the series regulator
and the function of the battery charger is to bring the nor the battery charger and becomes floating. It has
system operating point to the peak-power voltage by been shown in [2] that the operating point will move

86 IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS VOL. 35, NO. 1 JANUARY 1999
 solar array

Array

Peak Power
Discharger

R
...
r

: -. ..,
s
is

u Series
-+-
Regulator
E
.-
Tracker series
regulator
Vbus ref

solar array

Charger

-
series
regulator
voltage
oop, b
Fig. 12. Analysis at subsystem level.

focus on the design issues at each block rather than

Fig. 11. Control block diagram of RPFT system.
the system as a whole. Thus, identification of source
and load at the solar array interface and the bus
from the peak-power voltage to the desirable stable interface is important in order to formulate correct
equilibrium point near the solar array open-circuit small-signal models.
voltage (Fig. 10(b)), since the peak-power point is
on the right hand side of the separatrix. The solar
array operating voltage depends on the load demand. A. Eclipse Mode
Different loads will result in constant power load When the battery discharger regulates the bus
lines at different levels, and the intersection of each during eclipse mode, the small-signal model of the
load line with the solar array I-V characteristic as discharger can be formulated from the conventional
shown in Fig. 10(b) will determine the solar array boost small-signal model [6] with the appropriate load
operating voltage. If the load is increased such that small-signal model.
the equilibrium points B and C disappear, the system
has to switch to the PPT discharge mode , since the
system is not allowed to operate at the operating B. PPT Discharge Mode
point A. The battery charger must then be turned off,
To the discharger, the series regulator behaves
the series regulator switches to PPT, and the battery
as a load, as illustrated in Fig. 12(a), where v, is
discharger is turned on to regulate the bus voltage.
the bus voltage, and rL represents the simplified
load impedance. The series regulator provides
Ill. PROPOSED CONTROL SCHEME FOR THE RPPT constant peak-power at the bus side. Therefore, the
SYSTEM small-signal impedance of the series regulator seen by
the discharger is -R. Since id = i, - i,, then
A proposed control scheme for an RPPT system
is given in Fig. 11 where the structure of the control
scheme is in accordance with the functions of each
module as described in Table 11. Note that since
the battery charger and discharger functions are Assuming a constant power load (rL < 0), then, since
complementary, they can be combined into a single lrLl < R (because solar array capacity is less than
block (bidirectional buck-boost converter). Since the the load demand), re, will have an overall negative
battery discharger only needs to regulate the bus, impedance. This reduces to a conventional boost for
a voltage feedback loop L, is adequate to provide the discharger with a constant power load.
bus regulation. Loop Lrb of the series regulator As far as the series regulator is concerned, the
regulates the bus voltage during the PPT discharge load side becomes a stiff voltage source due to the
mode, whereas loop L,, regulates the solar array discharger regulating tlfe bus, and the controlling
voltage at the reference voltage, Vpeak generated by variable is the solar array voltage as illustrated in
the peak-power tracker. As far as the battery charger Fig. 12(b), where rs denotes the incremental resistance
is concerned, loop LCiis used for the TC mode, and of the solar array. This configuration is similar to
L , , regulates the solar array voltage at the peak-power the series-configuration PPT system [ 11. Thus the
voltage. small-signal model of the series regulator is similar to
the boost, but has a LHP zero instead of a RHP zero.
IV. SMALL-SIGNAL ANALYSIS.AND DESIGN
C. PPT Charge Mode
The most effective approach in small-signal
analysis and design of this RPPT system is to At the solar array interface v,, the battery charger
decouple the system into smaller subsystems and sees the solar array impedance r y and a constant power

HUYNH & CHO: DESIGN AND ANALYSIS OF A REGULATED PEAK-POWER TRACKING SYSTEM 87
 gain 1

'\ 1+T,+T

0 dB

Fig. 14. Reference-PPT-voltage-to-solar-array-voltage

transfer
Fig. 13. Small-signal block diagram of series regulator in PPT function.
discharge mode.
Peak-power tracker output reference
load from the series regulator, -R (Fig. 12(c)). Since voltage
i, = is - i,, then Gvbd Cycle-to-bus-voltage transfer function
1 - -Di, - -1 1 Gvsd Duty-cycle-to-solar-array-voltage transfer
(2) function
req Dv, rs R'
He Sampling effect at half of the switching
Thus the equivalent resistance seen by the charger frequency.
is a parallel combination of -rs/ - R. Because of the
peak-power tracker, the operating point can locate in The current loop I; and voltage loop are given
both current source and voltage source regions of the by
solar array, and consequently yS can assume a wide I; = FmR,G,,He (3)
range of magnitudes (note that r, is always negative). and,
Likewise, the load can change from no load to full T, = HvFmGvs, (4)
load, resulting in different -R. Therefore, req can
result in values from negative impedance to positive respectively. For optimum system performance in
impedance. Nevertheless, this configuration is similar current mode control, the voltage loop compensator
to the battery charger in the direct energy transfer is often of the form
(DET) system discussed in [3]. Thus, the small-signal
model of the charger is a boost type with a LHP zero (5)
instead of a RHP zero.
When the solar array output is regulated by the where w, is a low frequency zero which is used to
charger, it becomes a source to the series regulator. shape the system dynamic, and w p is a high frequency
Thus, during the PPT charge mode, the series pole. After closing both the current loop and the
regulator behaves as a conventional buck converter. outer voltage loop T,, the reference-PPT-voltage-to-
solar array voltage transfer function is given by
D. TC Mode HL Fm vsd

During the TC mode, the solar array moves into Grs= l+l;+r,
the voltage source region, and the small-signal models Since T, has a higher gain at low frequency and 2;
of the charger and the series regulator in this mode has a larger control bandwidth under standard design
are very much similar to those developed in the PPT practice, 1 + T, -I-7; can be approximated as in Fig. 14.
charge mode section. It follows that G,, from (6) has a flat dc gain at low
After analyzing the system mode by mode, it frequency, which is similar to the one developed for
is observed that the small-signal behavior of each the voltage mode control in [l]. Thus, it is expected
building block is similar to the conventional building that when closing the outer PPT loop, the resulting
blocks in the series-configuration PPT system and three-loop control system should be stable when the
the DET system. Thus, the d s i g n process can be sampling frequency, and step-size corrector for the
followed from the work done in [ l and 31. reference voltage are chosen appropriately as in [l].
However, some peculiarities observed from the The effect of continuous perturbation of the
RPPT system need to be analyzed. The additional reference voltage generated by the peak-power
PPT loop results in a three-loop control system as tracker on the bus regulation must also be analyzed.
current-injection control (CIC) is employed in each During the PPT discharge and the PPT charge modes,
building block [ 6 ] .Fig. 13 shows the standard CIC there exists a repetitive perturbation of the reference
small-signal block diagram of the series regulator
PPT voltage because the peak-power tracker has to
in the PPT mode, where the symbols are denoted as
update vpeak at discrete times. During the PPT charge
follows:
mode, the series regulator maintains bus regulation
VS Solar array voltage and the battery charger peak-power tracks. If the
'b Bus voltage series regulator is well optimized, resulting in low

88 IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS VOL. 35, NO. 1 JANUARY 1999
 PWM switch . . . . . . .
. . . . . . .
Gain

L
wz \ \

9Lr+ ,
0 dB \ frequency
G tb ,
\

w
; .
Fig. 16. Reference-to-bus voltage transfer function.
l:D
. . .
Fig. 15. Open-loop small-signal model of series regulator in PPT a fast battery discharger, the smaller the bus ripples
discharge mode. due to the effect of updating at the output of
the PPT becomes. A low-pass filter can also be added
closed-loop audiosusceptibility, it is expected that at the reference voltage input terminal to smooth out
continuous perturbation at the solar array output the discrete change of or even an additional bus
side does not affect the bus regulation. However, filter can be added at the load side.
during the PPT discharge mode, the series regulator
peak-power tracks, and the battery discharger
V. SlMlJLATlON RESULTS
regulates the bus. Since the battery discharger has
a finite control bandwidth, i.e. it takes a finite time A 3 KW RPPT system based on the proposed
for the discharger to regulate the bus under any control scheme in Fig. 11 is designed and simulated
disturbance, perturbation at the solar array output is using EASY5 simulator. Fig. 17 illustrates the circuit
likely to be observed at the bus side. Indeed, assume diagram of the RPPT system. The bus is regulated
that the battery discharger has some finite impedance at 120 V, the solar array peak-power voltage is
rb which is much smaller than the load impedance and approximately 160 V, and the battery voltage is set
Fig. 15 denotes the open-loop small-signal model at 100 V. Also, the sampling frequency of the PPT
of the series regulator, based on the pulsewidth controller is 1 KHz. Notice that the charger and the
modulated (PWM) switch model developed in [4], discharger are combined into a single bidirectional
where rs is the solar array incremental resistance; C buckhoost converter. The diode-ORs at the loop
and L are the series regulator input filter capacitance junctions; will select the lower control voltages.
and the series regulator inductance, respectively; The reference voltage of the series regulator in bus
D , 2 are the steady-state duty cycle and the duty regulation mode needs to be set slightly higher than
cycle perturbation; and ZL is the steady-state inductor that of the discharger to force the series regulator
current. The open-loop duty-cycle-to-bus-voltage into the PPT mode whenever the discharger regulates
transfer function (Fig. 13) is then given by the bus. This is done by setting the voltage dividers
that sense the bus to the opamp negative terminals of
both series regulator and battery discharger at slightly
different scales as shown in Fig. 17. Likewise, the
reference voltage generated by the peak-power tracker
is scaled down to slightly different voltages so that
(7) the battery charger is turned off whenever the series
regulator peak-power tracks the solar array power.
Notice that Gvbdhas a low frequency zero, and the de Fig. 18 illustrates the reference-PPT-voltage-to-
gain is controlled by r b ; the smaller rb is, the lower the solar-array-voltage transfer function (eq. (6))
dc gain becomes. After closing and T, as in Fig. 13, generated from the computer simulation. Notice
the reference-voltage-to-bus voltage transfer function that the magnitude curve has similar shape as
G,, is given by predicted in Fig. 14. The 34 dB difference in
the dc gain is due to the voltage divider in the
HiFm Gv bd resistive sensing network which has not been
Grb = 1 + 7; +. T, '
taken into account in the analysis. Fig. 19 shows
Low frequency and high frequency approximations the reference-PPT-voltage-to-bus-voltage transfer
can be performed, and the resulting asymptotic plot functions in PPT charge and discharge modes. It
of the above transfer function is illustrated in Fig. 16. is observed that the gain curve in the PPT charge
It follows that the battery discharger does not provide mode is much lower than that in the PPT discharge
as good bus regulation as the series regulator does mode. The small-signal simulation results imply
when the solar voltage is continuously updated by the that perturbation due to the PPT controller, which
peak-power tracker. Since the dc gain of (7) depends continuously updates the solar array voltage, is well
upon rb, this result shows that the smaller rb is, Le., attenuated in the PPT charge mode. Also, the shape of

HUYNH & CHO: DESIGN AND ANALYSIS OF A REGULATED PEAK-POWER TRACKING SYSTEM 89
 7 /
? I
__

.i
500UF 10 Ohms
47OuF_
-- Vload

solar array

5v

Fig. 17. Circuit diagram of RPPT for computer simulations.

Gain [dB] Phase [degree] Solar arrav voitaae M

122
Load voltage M
121 -
120-
119-
*
-\
Series regulator Inductor current [A]

Frequency [Hz]
Battery discharger inductor current [A]
20,
Fig. 18. Reference-PPT-voltage-to-solar-voltage
transfer function.

Gain [de] Battery charger inductor current [A]

20

10

0
0025 003 0035 0 04 0045 005
Time lsec]

Fig. 20. Transition from PPT discharge to PPT charge and back.

discharger at approximately 119.7 V. A ripple voltage

Frequency [Hz] of 0.4 V, which is due to the continuous update of the
Fig. 19. Reference-PPT-voltage-to-bus-voltage
transfer function. solar peak-power voltage at 1 KHz, is also observed
during the PPT discharge mode. Then, the bus voltage
switches to approximately 121.4 V after the transition
the gain curve in the PPT discharge mode agrees well to the PPT charge mode. As predicted from the
with the prediction (eq. (8)) shown in Fig. 16. analysis, because of the low audiosusceptibility of the
The simulation results in Fig. 20 show a transition series regulator, the low frequency voltage ripples are
from the PPT discharge to the PPT charge mode at well attenuated and becomes unnoticeable during the
30 ms and a backward transition to the PPT discharge PPT charge mode. Also, the oscillation around the
mode at 40 ms due to a step-change in the load peak-power voltage of 160 V in the simulation results
power. Notice that the bus is first regulated by the indicates that the solar array peak-power voltage is

90 IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS VOL. 35, NO. 1 JANUARY 1999
 the load, the battery charge current drops to zero
at approximately 50 ms, and the series regulator
switches 10 the PPT mode. The bus voltage is no
longer regulated by the series regulator and drops

:::I
118
Series regulator inductor current [A]
instantane:ously to 119.7 V as the discharger is turned
on.
4"

30
20
VI. CONCLUSIONS
10

The proposed RPPT system has been demonstrated

:I
0
Battery discharger inductor current [A]
to be effective for spacecraft power systems,
I ,, ,r r,, specificallly for LEOS, where it is more efficient than
0
Banery charger inductor current [A]
its series rand parallel counterparts. The proposed
control scheme is simple-there is no need for any
10 complicated central control unit for the system-and
0
the same amount of hardware is required; therefore,
0.025 0.08
0'04 Time [sec] O6 its applicability is very desirable. Large-signal and
Fig. 21. Transition from PFT to charge to TC and to PPT small-signal stability are analyzed. The results
discharge. conclude similarities among building blocks of this
system, the DET system, and the series-configuration
tracked. It is also observed that during the PPT charge PPT system.
mode, the discharger inductor current quickly drops to
zero, while the battery charger inductor current rises REFERENCES
as the battery charger regulates the solar array voltage. [l] Huynln, P., and Cho, B. H. (1992)
Fig. 21 shows the transition from the PPT charge Analysis and design of a microprocessor-controlled
to the TC mode at 30 ms and to the PPT discharge peak-power tracking system.
Presented at the Intersociety Energy Conversion
mode at 45 ms. As predicted from the large-signal
Engineering Conference, 1992.
analyses, the transition from the PPT charge to [2] Cho, :B.H., Lee, J. R., and Lee, E C. (1987)
the TC mode at 30 ms (due to a step-change in Large-signal stability analysis of spacecraft power
xcchrge in the battery charge current loop) causes the systems.
In Proceedings of IEEE Power Electronics Specialists
solar array output voltage to become floating and
Ctonference, 1987, 289-294.
move toward the voltage source region of the solar Kim, S. J., and Cho, B. H. (1990)
[3]
array. Also, during and after the transition, the bus Analysis of spacecraft battery charger systems.
voltage remains regulated at 121.4 V. Notice that the In Proceedings of the Intersociety Energy Conversion
transition to the PPT discharge mode at 45 ms due to Eiqgineering Conferrnce, 1990, 365-372.
a step-change to a heavier load causes a large droop [4] Voperian, V. (1990)
Simplified analysis of PWM converters using the model
in the solar array output voltage. This is because of the PWM switch, Pt. I and Pt. 11.
stepping the load demand to a larger load (this is IEEE Transactions on Circuits and Systems, 26, 3 (1990),
seen at the series regulator inductor current as it 490-505.
increases to a larger current at 45 ms) requires the [5] Chetty, P. (1987)
solar array to source instantaneously an amount of Ellectrical power system for low Earth orbit spacecraft
applications.
power more than its maxirnum capacity. As a result, In Proceedings of the Intersociety Energy Conversion
as shown in Fig. 10(b), the equilibrium points B Engineering Conference, 1987, 28 1-287.
and C disappear as the constant power curve moves [6] Sable, D., Ridley, R., and Cho, B. H. (1990)
upward, and the only equilibrium point of the system Cmomparison of performance of single-loop and
is A. Consequently, the system operating point must current-injection-control for PWM converters which
operate in both continuous and discontinuous modes of
move toward the current source region of the solar operation.
array I-V curve, implying that the solar voltage In. Proceedings of IEEE Power Electronics Specialists
must decrease as observed in the simulation results. C,unference(PESC), June 1990, 74-79.
Since all of the solar array power is used to source

HUYNH & CHO: DESIGN AND ANALYSIS OF A REGULATED PEAK-POWER TRACKING SYSTEM 91
 Phuong Huynh received his B.S. in 1990 , M.S. in 1992, and Ph.D. in 1994 from
Virginia Polytechnic Institute and State University, Blacksburg, all in electrical
engineering.
From 1995 to 1997, he worked for Philips Research as a Senior Member
of Research Staff, where he was involved in modeling and developing
high-frequency electronics ballasts for high-pressure discharge lamps. He is now
a Senior Member of Technical Staff at Cadence Design Systems in charge of
analog IC design for RF communications. His main research interests include
modeling, analysis and design of power electronics circuits, analog integrated
circuits, delta-sigma data converters, phase-locked loops, and RF communication
systems.

Bo H. Cho (M’89-SM’95) received the B.S. and M.S. degrees from the
California Institute of Technology, and the Ph.D. degree from Virginia
Polytechnic Institute and State University, Blackburg, all in electrical engineering.
From 1980 to 1982, he worked for two years as a member of Technical Staff
of Power Conversion Electronics Department, TRW Defense and Space System
Group. From 1982 to 1995, he has been a faculty member in the Department
of Electrical Engineering, VPI&SU, Blacksburg, VA. In 1995, he joined the
Electrical Engineering Department of Seoul National University, Seoul, Korea,
where he is currentiy an Associate Professor. His main research interests include
modeling, analysis, and control of power electronics circuits and systems, high
frequency power conversion, spacecraft power processing systems, and distributed
power systems.
Dr. Cho was a recipient of the 1989 NSF Presidential Young Investigator
Award.

92 IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS VOL. 35, NO. 1 JANUARY 1999