2
33 www.powersystemsdesign.com CONSUMER ELECTRONICS Frequency Compensation in Switching Regulator Design resulting frequency response contains a at area between the zero and the pole. Type-II compensation is generally used in applications where the output lter exhibits a single-pole roll-off at the cross-over frequency. The desired -20dB/decade roll-off at fc is achieved by ensuring that cross-over occurs somewhere in the at part of the error amplier’s response.  Part 2: Feedback path compensation By Nigel Smith, Business Development Manager, Portable Power, Texas Instruments In part one of this two-part series, the forward path of a switching converter was considered. In this  second and nal part, the feedback path is considered as the loop is closed and the overall circuit is compensated. O nce the gain and phase re- sponse of the forward path is known, the error amplier response can be designed. The main aims of frequency compensation are to ensure: (a) an adequate phase margin (typically >45°); and (b) an adequate gain margin (typically >10dB). In addi- tion, the loop gain should pass through unity with a slope of -20dB/decade. Before the frequency compensation can be designed, a suitable crossover frequency f c must be chosen. Switching converters with high crossover frequen- cies respond more quickly to changing operating conditions, and are therefore generally preferred; however, sampling theory limits the maximum crossover frequency that can be used. In practice f c typically lies between one tenth and one sixth f  sw , however, if the error ampli- er’s open-loop gain is insufcient at this f requency , f c may have to be further reduced. Having selected the desired crossover frequency, the gain, phase and slope of the forward path at f c can be obtained from its Bode plot. The required gain, phase and slope of the compensated error amplier at f c can now easily be obtained by comparing the two. Three types of compensation scheme are commonly-used, known as Ty pe-I, Type-II and Type-III (see Figure 1). Type-I is not commonly used in switching regu- lator circuits and will not be discussed here. T ype-II compensation exhibits a pole at the origin (to achieve high DC gain) plus an additional zero and pole. The Figure 1. Commonly Used Compensation Circuits and Their Response. SWITCHING REGULATORS

Switching Regulators Oct07

  • Upload
    mwisnia

  • View
    218

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Switching Regulators Oct07

8/3/2019 Switching Regulators Oct07

http://slidepdf.com/reader/full/switching-regulators-oct07 1/2

CONSUMER ELECTRONICS

Frequency

Compensation in Switching

Regulator Design

lti f t i

 Part 2: Feedback path compensation

By Nigel Smith, Business Development Manager, Portable Power, Texas Instruments

In part one of this two-part series, the forward path of a switching converter was considered. In this

 second and nal part, the feedback path is considered as the loop is closed and the overall circuit is

compensated.

Oth i d h Th t f ti h

SWITCHING REGULATORS

Page 2: Switching Regulators Oct07

8/3/2019 Switching Regulators Oct07

http://slidepdf.com/reader/full/switching-regulators-oct07 2/2

res lting freq enc response contains

Once the gain and phase re Three t pes of compensation scheme

SWITCHING REGULATORS

Power Systems Design Europe October 2007 34 35www.powersystemsdesign.com

SWITCHING REGULATORS

•Placethecompensationcircuit’s

zero(s)approximatelyoneoctavebelow

the output lter’s break frequency, and

calculate the value of K 1.Thisapproach

is relatively conservative, but avoids

the possibility of conditional stability by

ensuringthatphasestayswellabove0°

belowf c.

•Determinethenecessaryerror

amplier gain at f c and calculate the re-

quired error amplier gain at the zero(s).

•Calculatethemaximumphaselag

through the compensation circuit and,

using Table 2 or 3, and calculate the

minimum value of K  2achievingthis

phase lag. Calculate the frequency of

thecompensationcircuit’spole(s)

using.

• Calculat e the individual compo-

nentvaluesinthecompensationcircuit

required to achieve this response.

Figure 2 shows typical Bode plots for

the forward path, error amplier and

overall response of switching converter

usingType-IIIcompensation.

Each engineer has his own preferred

approach toward frequency compen-

sation, and in practice some iteration

will usually be necessary, however, the

approach described above provides a

good starting point f or inexperienced

engineers to build a stable circuit with

adequate performance.

 A general procedure for compensat-

ingaswitchingconvertercannowbe

simplied to the following:

• Generate the forward-path Bode

plots.

• Select a suitable cross-over fre-

quency. Use the rule-of-thumb that f c

should lie somewhere between one

tenth and one sixth the switching fre-

quency, but may need to be reduced if

the error amplier’s open-loop gain at

this frequency is insufcient. Determine

the forward-path gain and phase at the

crossover frequency.

• From the slope of the forward-path

gainatf c, determine whether Type-II or

Type-III compensation is required. If the

forward path slope at f c is -20dB/de-

cade, then Type-II compensation should

be used; if the slope is -40dB/decade,

thenType-IIIisnecessary.

high frequencies and a

slope of +20dB/decade

in the middle of the fre-

quency range of interest.

Type-IIIcompensationis

typically used to com-

pensatecircuitswhere

the output lter exhibits a

double-pole at the

cross-over frequency.This is done by ensuring

thatcross-overoccurs

midway up the error

amplier’s +20dB/decade

slope; the combined ef-

fect of the error amplier

and output lter slopes

is the desired -20dB/de-

cade response.

Therelativeposition

of the poles and zeros in

thecompensationcircuit

determines the overall

phaseboostoccurring

atf c. Thus, by placing

the poles and zeros at

suitable frequencies, the

desired phase margin can

be achieved. There are

a number of ways to ap-

proachthis.Onewayisto

consider the position of

the low frequency zero(s)

and high frequency

pole(s) using two factors

K 1 and K  2, as follows:

By considering the rela-

tive values of K 1 and K  2,

thephaseboostat f ccaneasily be determined from

Tables 1 and 2.

The gain of a Type-II

compensationcircuitat f c

is equal to the gain AV at

thezero.Type-IIIcom-

pensationhasagainin

dB at f c given by:

G = AV1 + 20log(K 1 )

whereA V1 is the gain at the second zero

in dB.

Type-IIIcompensationcontainstwo

zeros and two poles in addition to a pole

attheorigin.Theresultingresponse

contains an area of increased gain at

www.ti.com

Table 1. Phase Change through a Type-II Compensation Circuit.

Table 2. Phase Change through a Type-III Compensation Circuit.

Figure 2. System Bode Plots.