AIC AN015 Dc/dc power converter Datasheet

AN015
The effect of ESR on DC/DC power converter
Introduction
The capacitors, inductors, and resistors in circuit
C=
εA
… … … … … … … … … … … .(equ.1)
d
analysis have ideal properties. Capacitors and
inductors follow precise derivative relationships. In
the real world, components do not follow such simple
models. The non-ideal behavior of energy storage
components is very important in the context of power
electronics. This article is to discuss the effect of
capacitor’s
ESR
(Equivalent
series
Resistor)
characteristics on DC/DC power converter. And now,
capacitors shall be presenting first.
where εis the electrical permittivity of the insulating
layer, A is the surface area, and d is the distance
between 2 plates. By definition, the magnitude of the
charge Q stored on either plates of the capacitor is
directly proportional to the potential difference V
between the plates. Therefore, we may write
Q = CV … … … … … … … … … … … . (equ.2)
where V is the applied voltage. In the arrangement of
parallel plates, the electric field is given by
Capacitors- Principles and Equivalent
Circuits
E=
V
… … … … … … … … … … … . .(equ.3)
d
The time derivative of Q=CV gives the standard
property
i =C
dv
… … … … … … … … … … … . ..(equ.4)
dt
The parallel plate arrangement has provided some
challenges for years. Picture that two plates with
measurements of one square meter for both plates,
Figure 1. The Basic Parallel-Plate Capacitor
Geometry[1]
Capacitor devices come in many forms, but nearly all
of them are composed of conducting plates or films,
also separated by an insulating layer such as air or
paper. A common parallel arrangement, shown in
figure 1, gives equation 1 as follows.
which are separated by 1um air space, form a
capacitor of less than 10µF. Nevertheless, typical
capacitors used in power converter are usually on the
order of 100µF or more. However, the permittivity of
free space or air is small, withε0 = 8.854pF/m and
very large plates are required to give significant
values of C. But the effect of inserting the dielectric is
Dec 2000
1
AN015
to increase the capacitance by a factor of dielectric
numbers of important properties.
constant k.
Ø The current is not exactly consistent with i=Cdv/dt
e
K=
… … … … … … … … … … … . (equ.5)
e0
Ø Current flows even when dc voltage is applied
Ø Resonance occurred due to the combination of L
Typical values of k are shown in table 1 as below.
Table1. Dielectric Constant and Strength
Material
Air
Paper
Glass
Paraffin
Rubber
Mica
Water
Dielectric
Constant
1.00059
3.7
4~6
2.3
2~3.5
6
80
Dielectric
6
Strength(10 V/m)
3
16
9
11
30
150
-
and C. The device behaves as an inductor when
operated at a frequency which is over the
resonant frequency.
Ø Extra power loss
The properties are out of consideration in general
applications. However, they play as a decisive role in
some of the applications for high operating frequency
or low voltage with high current, e.g. VRMx.x.
The inductance is expected to be in the nano-henry
Although the term “dielectric” is nearly invariable
applied to insulating materials, which is somewhat a
misleading. Water, which has a high dielectric
( 10 −9 H ) range, since it represents wire inductance
and geometric effect. The leakage resistance should
be very high, and the time constant of the insulation,
constant, is not an insulator.
τ = R LEAK C … … … … … … … … … … … . (equ.6)
In a capacitor device, the wires and plates have
should be very long.
resistance and inductance. The insulation is not
perfect and has a leakage resistance. These direct
In many power electronics applications, a capacitor is
properties can be summarized in a circuit model,
to be used at a specific frequency. Given a radian
shown in figure 2.[2]
frequency ω, the circuit model can be evaluated as
a set of impedance, then simplified. The simplified
one is shown in figure 3, the derivation starts with a
capacitor C in parallel with the leakage resistance
RLEAK.
Figure 2. General Circuit Model for a Real
Capacitor
The properties of the model can be minimized, but
not be avoided altogether. This equivalent exhibits
Figure 3.
2
AN015
widely used by manufacturers as the basis of
Fig. 3 can be turned into an equivalent series, as
figure 4, after some certain mathematical transform.
specifications. Since the ESR is obtained through a
transformation,
therefore,
it
is
a
non-linear
frequency-dependent resistance. The value is given
at a specific frequency (120Hz or 100KHz is common)
on data sheets.
Figure 4.
2
2
2
Since RLEAKC is high, ω RLEAK C >>1 is true as the
frequency
goes
above
about
reactance portion simplifies to
resistance portion simplifies to
1Hz.
Then
the
−j
. The series
ωC
1
ω 2RLEAKC 2
Figure 6. The standard model of a capacitor
. It gets
Besides, the parameter, dissipation factor df, is often
much “lovelier” for the exhibition in figure 5,
used to indicate the quality of a capacitor. It shows
comparing with figure 4.
the ratio of resistance to reactance. For the frequency
1
below
, the standard model’s reactance is
ESL C
approximately
df =
1
, which gives
?C
R
= (ESR)? E … … … … … … … … … … … . (equ.8)
X
Figure 5.
The ratio is also called loss tangent, df, coming with
Till now, the equivalent circuit has been simplified to
a series R-L-C combination. The inductor in the
circuit is termed as an equivalent series inductance
or ESL. The capacitor represents the internal ideal
capacitance effect. The resistor, termed as an
X
an impedance angle f = tan −1( ) . If the wire
R
resistance is small, the ESR becomes the second
term in equation 7 and the dissipation factor can be
written as
equivalent series resistance or ESR, has the value of
ESR = Rw +
1
2
ω R LEAK C
2
… … … … … … … . (equ.7)
The combination, shown in figure 6, is sometimes
df = tan δ = ESRωC ≒
ωC
2
ω RLEAKC
2
=
1
ωRLEAKC
… .. (equ.9)。
called as a standard model of a capacitor. It has been
3
AN015
The leakage resistance can be calculated from the
process. Even when LDO used, the ESR effect
resistivity of insulation, ρ ,since the geometry is
appears as the variation loading current applied.
known. We substitute
εA
ρd
and
for C and RLEAK
d
A
and derive the result tan δ =
1
from equation 9.
ωερ
Tutorial example
Switching mode power supply part
Thus the loss tangent is independent of geometry
and
can
be
considered
a
material
property
As the IT market grows rapidly, the device power
associated with the insulator. That is, the ESR value
requirement is getting more and more complicated.
strongly depends on the choice of insulation
Obvouusly, one trend is to use the switching mode
material.
power supply instead of conventional linear regulator,
which betters the efficiency.
The effect of ESR
This design example, as showm in figure 8, provides
Unfortunately, the effect of ESR will never disappear.
Below the resonant frequency, a real capacitor will
show a resistance in series with a capacitive
reactance. The capacitor becomes a simple RC
series combination. In some applications such as a
an output voltage of 1.3V to 3.5V, which meets the
VRM8.4 specification. It can deliver high current to a
CPU VCORE load. According to the tight VRM
transicent demand, the effect of ESR plays a
significant role in the application.
DC/DC converter, a capacitor is often exposed to a
square wave of current. Ideally, this produces a
triangular voltage across the part shown in figure 7.
Figure 7.
But to lead to some unusual result, the ESR will
exhibit an ESR voltage drop. With the ESR, a small
square wave adds in series with the triangle, the
abrupt voltage change called an ESR jump. In some
case, it can dominate over ripple during the design
4
AN015
AIC1570- Advanced PWM Switching and Dual Linear Controller IC!!!
The AIC1570 combines a synchronous voltage mode controller with a low dropout linear regulator and
a linear controller as well as the monitoring and protection functions in this chip. The PWM controller
regulates the microprocessor core voltage with a synchronous rectified buck converter. The linear
controller regulates power for the GTL bus and the linear regulator provides power for the clock driver
circuit.
An integrated 5 bit D/A converter that adjusts the core PWM output voltage from 2.1V to 3.5V in 0.1V
increments and from 1.3V to 2.05V in 0.05V increments. The linear regulator uses an internal driver
device to provide 2.5V±2.5%. The linear controller drives with an external N-channel MOSEFET to
provide 1.5V±2.5%.
Features
l Provides 3 Regulated Voltages for Microprocessor Core, Clock and GTL Power.
l Simple Voltage-Mode PWM Control.
l Dual N-Channel MOSFET Synchronous Driver.
l Operates from +3.3V, +5V and +12V Inputs
l Fast Transient Response.
l Full 0% to 100% Duty Ratios.
l ±1.0% Output Voltage for VCORE and ±2.0% Output Voltage Reference for VCLK and VGTL.
l TTL Compatible 5-bit Digital-to-Analog Core Output Voltage Selection. Range from 1.3V to 3.5V.
l 0.1V Steps from 2.1V to 3.5V.,0.05V Steps from 1.3V to 2.05V.
l Adjustable Current Limit without External Sense Resistor.
l Microprocessor Core Voltage Protection against Shorted MOSFET.
l Power Good Output Voltage Monitor.
l Over-Voltage and Over-Current Fault Monitors.
l 200KHz Free-Running Oscillator Programmable up to 350KHz
5
AN015
C18
+12VIN
R15
VCC
10
1000pF
R2
1
20
C16
1µF
24
23
L1
OCSET 2.2K
C15
Q1
UGATE
+
1µH
C1-C7
6 x 1000µF
1uF
PHASE
+5VIN
GND
VOUT1
L2
+3.3VIN
VIN2
+
C19
1000µF
12
22
Q3
GATE3
VOUT3
R11
FB3
21
+
C24-36
15
16
2.5V
3.5µH
Q2
LGATE
D5820
PGND
R4
4.99K
7 x 1000µF
10K
+
R12
10K
C43-46
4 x 1000µF
19
VSEN
C40
VOUT1
waveform
shown in
figure 9
0.68µF
VOUT2
VOUT2
R8
2.21K
FB1
13
18
1.5V
VOUT3
waveform
shown in
figure 13
+
C47
R13
1.87
K
FB2
270uF
R10
160K
C41
10pF
R9
732K
11
C42
R14
10K
2.2nF
17
7
PGOOD
FAULT
VID0
6
8
VID1
5
10
VID2
4
VID3
3
VID4
2
9
14
COMP1
RT
SS
C48
40nF
Figure 8. AIC1570 3 in 1 Power Solution for Motherboard[3]
Figure 9.VOUT1 of AIC1570
6
AN015
Generally, we place several single capacitors in
The circuit is initiated at t=t0, during the rising or
parallel, which reduce the total ESR. And figure 9
falling edge of the CPU supply current transient. A
shows the real output voltage waveform when
voltage drop is produced at t=t1 due to influence of
constant current pulled. It still dominates a certain
ESR and ESL, which gives
percentage of the output variation although the effect
of ESR jump is slight. Also, the waveform would be
VDROP@t1= ESR × ∆IOUT + ESL
∆IOUT
… … (equ.10)
∆t
worse if neither the measurement is taken skillfully
enough nor appropriately layout applied.
After the initial transient, the ESL term in equation
10 drops off as in equation below
Figure 10 shows a 35A load current occurred in about
VDROP@t2= ESR × ∆IOUT … … … … … … … . (equ.11)
8A/us current slew rate. And figure 11 is a redraw
waveform from figure 10, which presents clearer
Within a clock cycle of the output current transient,
the current feedback loop detects change of output
transient response.
current and increases power-switch duty cycle to
maximum. Figure 12 shows the transient step and is
followed by the description.
IL
IOUT
IL
(0)
IOUT↑
(1)
IOUT↑-IL
IL↑
Figure 10.
IOUT↑
(2)
IL↑
(3)
IOUT↑-IL↑
IL↑
IOUT
IOUT↑
IL↑-IOUT↑
IOUT↑
(4)
VOUT
Figure 12.Transient Current Flow Sequence
t0
t1 t2
t3
t4
Figure 11.
7
AN015
Figure 13 illustrates the output drop due to the
(0) Circuit is initiated at t=t0
unavoidable ESR and ESL effect, which follows the
(1) Capacitor provides most of transient current while
output current transient occurs
equation 11 and 12. And we can see the difference
from the implementation of the larger and smaller
(2) The output inductor current starts to ramp up
output capacitors at the same test condition.
(3) Inductor current goes up to cover both of the
output current and the current charge back to
Conclusion
capacitor
(4) Output current is provided by inductor current
The following tips should be noticed to prevent your
applications from ESR and ESL effects.
l
Linear output part
Care must be taken for the different placement as
shown in figure 14.
l
An appropriate layout applied.
l
Sufficient grounding implemented.
l
Suitable measurement taken as in figure 15.
[4]
Figure 13. Output Transient in LDO Controller Output
8
AN015
figure 14. The Placement of Capacitors[5]
Figure 15. Correct Measurement for Transient Response
Reference
[1] Harris Benson, University Physics, John Wiley &
Sons, 1995, Chapter 26
[2] Philips T. Krein, Elements of Power Electronics,
Oxford Univ Press, 1998, pp384-385, 392
[3] Analog integrations Corporation, AIC1570 Data
Sheet, AIC, 2000
http://www.analog.com.tw/products.htm
[4] Gloria, The best source of Electrolytic Capacitors,
GAE Radial low ESR,
[5] http://www.glorai-cap.com.tw
[6] Gloria, The best source of Electrolytic Capacitors,
General Information for Application,
http://www.glorai-cap.com.tw
9
Similar pages