Structure of LDO

Download Report

Transcript Structure of LDO

線性穩壓器 (2)
Linear Regulators (2)
Instructor: Po-Yu Kuo (郭柏佑)
國立雲林科技大學
電子工程系
Structure of LDO

A typical series regulator which consists of four main
building blocks:
Vin
Vref
+
A(S)
Vn1
Power
Transistor
-
Vo
R1
Re
Vfb
R2
RL
Co
2
Structure of LDO

Voltage Reference (Vref): a very stable voltage with respect
to temperature change and input voltage variations, usually
of the bandgap type.

Error Amplifier (A(s)): a very high (dc) gain opamp to
achieve a close to zero error signal Verr=V+ - V-.

Feedback Network: R1 and R2 define the feedback factor
and generate Vfb to be compared with Vref to get the
designed output voltage Vo.

Series Pass/Power Transistor (Q1): power transistor
configuration to pass high current from the source to output.
As it handles large current, the size of pass transistor
dominates the area of the whole series regulator.
3
Structure of LDO

Dropout Voltage (Vdo) is the
minimum voltage difference
between the input and output
under which the regulator still
able to maintain the output
within the specification.
Vin
Vref
+
A(S)
Vn1
Power
Transistor
-
Vo
R1

With a Li-Ion battery as Vin,
Vin varies from 2.7V to
4.2V.
Re
Vfb
R2
RL
Co
Vdo,max =4.2-Vov,ML
4
Structure of LDO
5
Specifications of LDO

Two Categories:

Regulating (accuracy) performance
Line regulation, load regulation, temperature dependence,
transient overshoot, transient recovery time, stability
Power Characteristics
Io, Quiescent current Iq, Vin & Vo (

)
6
Efficiency



Current Efficiency  I :
I  I o /( I o  I q ) where Iq is the quiescent current of LDO
In LDO design for portable applications, Io is usually
much larger than Iq with > 99% efficiency
When Io is 0, Iq should be minimized (I3 should be small
by large values of R1 and R2)
7
Efficiency

Smaller dropout voltage causes a higher power
conversion efficiency especially Io >> Iq

In light-load condition (small Io), the efficiency is poorer
as I1, I2, and I3 are close to Io
8
Dropout Voltage and Power-Transistor Sizing





VSD must be always larger than Vov at different conditions
Design at the worst case: largest Vov at Io(max) and μp(min) at
the maximum temperature
By using minimum L (the smallest transistor and hence
parasitic capacitance), keep increasing W until meeting the
dropout specification
IR at the routing metals increase VDO
Design margin by experience-generally the chosen W is 1.11.2 times of the theoretical W
9
Load Regulation




Load Regulation (R): closed-loop output resistance of
LDO
Ro is the open-loop output resistance of the pass
transistor as Rf1, Rf2 >> Ro
Better load regulation is achieved by smaller Ro (using
minimum channel length of the pass transistor) and
larger loop-gain magnitude
As Ro
1/Io, high Io range gives better load regulation
10
Line Regulation




gmp is the transconductance of power PMOS transistor
Line regulation is independent of the gain of the power
transistor
Line regulation can be improved by a high-gain error amplifier
Other error sources on line regulation are voltage reference
and offset voltage of the error amplifier
11
Review on Voltage Gain


Gm and Ro can be found individually
Input-Output voltage gain can be found by the product of
Gm and Ro
12
Line Regulation Including Other Errors


Voltage gain of the error amplifier is not the only
parameter to improve line regulation
Good designs on supply independence of Vref and
reducing systematic offset of error amplifier are very
important
13
Temperature Coefficient


Variation of Vo at different temperature depends on both
voltage reference and error amplifier design
Rf1 and Rf2 must be made by the same material and
closely placed
14
Load Transient Response
15
Load Transient Response

Better load transient response by tresp ↓, Co ↑, Re ↓ and
Lc ↓ .
16
AC Design (1): Loop-Gain Analysis
17
AC Design (2): Loop-Gain Analysis
18
AC Design (3): Loop-Gain Analysis
19
AC Design (4): Loop-Gain Analysis





ze should cancel p2 within one
decade of frequency for stability
Parasitic pole(s), ppar, must be far
away from the unity-gain
frequency (UGF)
Different UGFs are resulted from
different Re values such as ze
locating before or after p2
p2 locates at very low frequency
as Cpa and ra are large
Required large Co and Re


Large Co is unfavorable in the cost
consideration
Low-frequency pole-zero cancellation
is unfavorable to load transient
recovery time
20
LDO with Voltage Buffer



Smaller required Re can be achieved by inserting a low
output-resistance (1/gmb) voltage buffer
One more pole (p3) is created but is located at high
frequency due to small output resistance of the voltage
buffer
p2 (with voltage buffer) locates at a higher frequency than
the one without voltage buffer (Cb << Cg)
21
Effect of Load Currents on Stability

Loop gain is larger
when Io is smaller due
to gmpro 1/ √Io

p1 is lower when Io is
smaller due to larger ro
of the power transistor
(ro 1/Io)

Worst-case stability at
maximum Io

Compensation at max.
Io
22
Effect of Loop-Gain Magnitude on Stability





Larger loop gain by
increasing ra of the error
amplifier
p2 → p2’
A larger Re is needed to
create a zero at lower
frequency
(ze → ze’)

Larger loop gain → more
unstable as p3 may be
below the UGF of loop
gain
A larger Co is generally
needed
23
Loop Gain Simulation
24
Summary of LDO Specifications
25
Circuit Implementations

Circuit of LDO consists of







R1 and R2
Cin and Co
Vref
Error Amplifier
Voltage Buffer
Power Transistor
Vin,min = Vov,Me1 + Vgs,Mb2 + Vgs,Mp Low-voltage operation impossible!
26
Circuit Implementations




BJT has a small VBE drop (~0.7V)
The circuit can operate at lower input supply compared
to the previous case
Smaller input capacitance for small VBE
Base current introduces larger offset voltage and hence
degrades accuracy of the output voltage
27