State Space Averaging Technique of Power Converter

with Digital PID Controller

M. F. N. Tajuddin, N. A. Rahim, I. Daut, B. Ismail and M. F. Mohammed
School of Electrical System Engineering
Universiti Malaysia Perlis
01000 Kangar, Perlis, Malaysia

Abstract—This paper describes a digital Proportional-Integral- resistors, capacitors, inductors and operational-amplifiers is
Derivative (PID) controller technique applied to the DC-DC buck used for implementation of controlling algorithm, while in
converter. The converter is modelled using a state space digital control, microprocessors are used and controlling
averaging technique. Due to the non-linear of the power algorithm can be programmed and feed to microprocessor.
converter, the PID controller is designed to simplify the Digital control system offer many advantages over their analog
compensation of the DC-DC converter. The control algorithm is counterparts [5]. An important advantage offered by digital
implemented in Digital Signal Processor (DSP) TMS320F2812. control is in the flexibility of its modification controller
The steady state and dynamic response performances of the characteristics, or of adopting the controller if plant dynamics
controller are demonstrated. Experimental results are provided
change with operating conditions. The ability to redesign the
to demonstrate the effectiveness of the design.
controller by changing the software (rather than hardware) is an
Keywords-State-space averaging technique; digital PID important feature of digital control as against analog control.
controller; buck converter Furthermore microprocessors are much less prone to
environmental conditions than capacitors, inductors, etc. The
I. INTRODUCTION programs can scale to the limits of the memory or storage space
Switching-mode converters are widely used today to without extra cost and parameters of the program can be
provide power processing for applications ranging from changed with time. This paper demonstrated the effectiveness
computing and communications to medical electronics, of the digital PID control applied to the state space averaging
appliance control, transportation, and high-power transmission. DC-DC converter model. The control algorithm is
Originally distributed regulated converters were used in implemented in DSP TMS320F2812.
mission-critical space applications, where reliable operation II. MODELLING OF DC-DC SWITCHING POWER
was of prime concern [1]. These converters are nonlinear CONVERTER
dynamical systems. The nonlinearities arise primarily due to
switching, power devices, and passive components, such as The inherent switching operation of power electronic
inductors, and parasitic. The static conservation properties of converters results in the circuit components being connected
the elementary switching converters (buck, boost, and buck- together in periodically changing configurations, each
boost) have been thoroughly understood since the early 1970s. configuration being described by a separate set of equations.
This is one of the main reasons of their ever-increasing number The transient analysis and control design for converters is
of applications in electrical energy conversion. Modeling and therefore difficult since a number of equations must be solved
analysis of switching DC-DC converters can be either in sequence. The technique of averaging provides a solution to
numerical or analytical. In numerical techniques, various this problem. A single equation may be formed to describe the
algorithms or circuit simulators are used to produce converter approximately over a number of switching cycles by
quantitative results. These methods are easy to use. They simply taking a linearly weighted average of the separate
posses accuracy and universality and they are applicable when equations for each switched configuration of the converter.
no equivalent model is available. However, they are time- State space averaging [3, 6] is the most common averaging
consuming and fail to provide the design insight needed to technique, and is used here to model the switching DC-DC
understand the behavior of switching regulators. Analytic converter.
techniques, on the other hand, provide analytic expressions A. Averaging Small-Signal Modelling
representing the operation and performance of the converters.
Perhaps the most popular continuous-time technique is the In state-space averaging (SSA), the switching circuit is
small-signal analysis, which uses either, circuit averaging [2], divided into two (CCM) or three (DCM) different structures.
state-space averaging [3], or PWM switch modeling [4]. The derivatives of inductor currents and capacitor voltages are
defined based on circuit theory for every substructure. These
The PID control method can be implemented digitally or currents and voltages are averaged over one switching cycle.
analog to control parameters of converter such as voltage, The state space representation for a voltage-mode controlled-
current, etc. In analog control, analog components such as buck converter shown in Fig. 1 will be derived next.

978-1-4244-4547-9/09/$26.00 ©2009 IEEE TENCON 2009

1
 Rearranging (1c) yields:

(1d)

Substituting in (1d) into (1a) and (1b) yields:

1
(1e)

1
(1f)

The voltage across the diode is equal to zero when the

MOSFET is off. Therefore during, the circuit in Fig. 1(b)
can be used as a model of the Buck converter. If, is equal
to zero, the circuit in Fig. 1(a) and Fig. 1(b) are actually the
Figure 1. Example Buck Converter with Digital Controller same. As a result, a state space model when the switch is off
can be carried out by setting equal to zero in (1e). By
setting equal to zero yields:

(2)

The inductor current, and capacitor voltage, are

chosen as the state variables. is chosen as an input signal
while as the output signal. Using Equations (1d), (1e) and
(1f), the following state space system is obtained during :

Figure 1(a). The circuit of Buck converter during,

(3)

where

(4)
+

(5)

(6)
1

Figure 1(b). The circuit of Buck converter during,

During the on time, the switch is on, and the diode is off. 1⁄ 0 (7)
Therefore, the corresponding sub-circuit is as shown in Fig.
1(a). Applying Kirchhoff's Voltage Law (KVL) and Kirchhoff's
Current Law (KCL) the following equations are obtained.
1 (8)
(1a)
Using (1d), (1f) and (2), the following state space system is
1 obtained for the circuit in during off time:
1(b)

(9)
1(c)

TENCON 2009

2
 where
(10)
0 0 (11)
(20)
(12)
The averaged state-space equations can be obtained by
multiplying (3) and (9) by the duty cycle, and 1
respectively, and the resulting equations are added together as
follows
1
(21)
(13)

1 where
By taking the average of the two linear equations, a set of (22)
nonlinear continuous equations (13) is obtained. This equation
is an approximation of the time-variant system. (23)
To study the small signal behaviour, the time varying (24)
system described in (13) can be linearized using perturbation Assuming that departures of the states from their
technique. Small ac perturbations are introduced in the dc corresponding steady state values are small compared with the
steady state quantities. Capital letters denote the operating- steady value themselves allows the second order terms in (22)
point (dc, steady-state) values and the hat-symbol (^) denotes to be neglected. The averaging model in its final form is shown
perturbation (ac) signals. below.
0 (25)
(14)
(26)
(15)
The ac model:
(16)

(27)
In general . However to obtain the
transfer function between output voltage, and duty
ratio, , the perturbation is assumed to be zero in the
(28)
input voltage to simplify the analysis. Therefore,
Equations (25), (26, (27) and (28) represent the small-signal
(17)
low frequency model of any two state switching DC to DC
Substitution of (14)-(16) into (13) yields: converter working in the continuous conduction mode.
B. Transformation of AC Equations into s-Domain to Extract
the Transfer Functions
Equation (27) and (28) consists of the ac perturbation.
(18) Using laplace transformation in (27) and (28),

(29)

(19)
(30)
Grouping in Equation (29) yields:
Small signal analysis starts by recognizing that the
(31)
derivative of steady state component, is zero. By collecting
term

TENCON 2009

3
 Figure 2. Block Diagram of Buck Converter with PID controller

where I is a unity matrix. Substituting x ̂(s) in (30) results in IV. IMPLEMENTATION OF DSP BASED PID CONTROLLER
The PID algorithm can be expressed in the discrete-time
domain as:
(32)
1 (35)
Therefore, the transfer function of output to variation in the 1
duty cycle is expressed as
where
is the proportional mode control gain
(33)
is the integral mode control gain
is the derivative mode control gain
Substituting (6)-(12) into (33) yield the important small signal Time relationship:
model for DC-DC Buck converter that will be used in
designing the controller. is the sampling period
is the discrete-time index: = 0, 1, 2, …
1
(34)
1 1

III. SYSTEM CONTROL To eliminate the need to calculate the full summation each
time step (which would require an ever increasing amount of
The feedback controller used in this application utilizes the computation as time goes on), the summation is expressed as a
PID algorithm as shown in Fig. 2. The converter output running sum:
voltage, is sensed and fed back to a comparator which
compares with reference voltage, . The instantaneous 1 1 (36)
current error , is fed to a PID controller. The integral term in
the PID controller improves the tracking by reducing the
instantaneous error between the reference and the actual (37)
current. The resulting error signal is compared with a 1
triangular carrier signal and intersections are required to
The flow chart in Fig. 3 shows steps involved in
produce PWM signals for the inverter switches [7].
implementing digital PID control in DSP TMS320F2812.

TENCON 2009

4
 Figure 3. Flowchart of digital PID control in TMS320F2812

V. EXPERIMENTAL RESULTS A. Open Loop Response

In the open loop system the voltage overshoot under load
The proposed digital PID method is implemented in the
can be unsatisfactory. Fig. 4 shows the voltage response under
DSP TMS320F2812. Table I shows the specifications of the
open loop control during start up. It shows there is an
DC-DC Buck converter implemented in this work. These
oscillation and high overshoot in the voltage and current
parameters are calculated on the basis that the converter is to be
response in spite of the damping provided by the load
operated in CCM for 20 V to 30 V input supply at 0.5 A to 2 A.
resistance. It is also seen that the steady state error exist in the
The maximum peak to peak ripple voltage is 50 mV.
open loop system.
TABLE I. SPECIFICATION OF DC-DC BUCK CONVERTER
Parameter name Symbol Nominal Value / Spec. no
30V
Input voltage
20V
Output voltage 12V
Output Capacitor 220μF
Inductor 2.12mH 1>

Load resistance 18 Ω
Switching
20kHz
frequency
1) Ch 1: 5 Volt 25 ms
Power MOSFET M IRF 530
Figure 4. Open loop start up transient response of Buck Converter
Power diode D MUR 820

TENCON 2009

5
B. Closed Loop Response VI. CONCLUSION
PID controller improves the start up transient response This paper describes a Digital PID voltage-mode control
performance of the converter. Fig. 5 shows the voltage method for DC-DC Buck converter. Although many voltage
response under feedback control during start up. It can be seen controlling methods are available but in fact, the PID controller
that there is no overshoot and that the settling time is improved. is difficult to outperform for several reasons [8]. The integrator
The system is more stable and able to maintain a constant increases the system type number, thus minimizing steady-state
output voltage. error. Two zeros in the controller make it possible to dampen
resonant characteristics and improve transient response.
Compared to many other control philosophies, the PID
controller structure is fairly easy to explain and understand.
To maintain its reliability, digital method was chosen since
it has many advantages compared to analog method. It provides
robust controlling method for the converter. To smooth the
1> progress of implementing DSP based PID controller, the design
guidelines for the controller is described in a flow chart.
Experimental waveforms obtained have been used to validate
the digital PID control method and its feasibility using DSP.

1) Ch 1: 5 Volt 5 ms REFERENCES
Figure 5. Start up transient response of Buck Converter with digital PID
control [1] Cho, B. H. (1985). Modeling and Analysis of Spacecraft Power Systems,
PhD Thesis. Virginia Polytechnic Institute and State University. 181 p.
Fig. 6 displays the output voltage and current waveforms of [2] Wester, G. W. and R. D. Middlebrook (1972). Low-frequency
the converter operating with the input voltage of 24V at half characterization of switched dc-dc converters. IEEE Transactions on
Aerospace and Electronics Systems 9(3), 376-385.
load. It can be proved that the system with PID controller
[3] Middlebrook, R. D. and S. Cuk (1977). A general unifed approach to
produces the required output voltage and current. These results modelling switching-converter power stages. In: Proceedings of Power
are in well agreement with the theoretical values. Electronics Specialists Conference. pp. 521-550.
[4] Vorperian, V. (1990). Part I: Continuous conduction mode, simplifed
analysis of PWM converters using model of PWM switch. IEEE
Transactions on Aerospace and Electronics Systems 26(3), 490-496.
[5] B. Duan, Y. and Jin, H. (1999). Digital controller design for switch
mode power converters. Conf. Procs. of the Fourteenth IEEE Applied
Power Elect. Conf. and Exposition. 2, 967 - 973.
[6] Mohan, N., T.M. Undeland and W. P. Robbins (1995). Power
electronics: converters, applications and design. John Wiley & Sons,
1> New York, 1995.
[7] David M.Alter.: “Thermoelectric cooler control using TMS320F2812
DSP and a DRV592 power amplifier” Texas Instruments Application
Report.
2>
2 [8] Y. Feng, Q., Nelms, R. M. and Hung, J. Y. (2006). Posicast-based digital
control of the buck converter. IEEE Trans. on Indust. Elect.. 53(3), 759 -
1) Ch 1: 5 Volt 25 us 767.
2) Ch 2: 2 Volt 25 us

Figure 6. Waveforms of output voltage and current at steady state operation

under nominal operating conditions

TENCON 2009