Download Document

Optimizing Efficiency of Switching Mode Chargers
Multi-Cell Battery Charge Management (MBCM)
TI Information – Selective Disclosure
Outline and Purpose

Understand the key parameters of a MOSFET and the relationship
to power loss of a switching charger
1.
Conduction loss
2.
Switching loss
3.
Gate drive

Inductor selection and its impact to the loss

Current sensing resistance vs. the loss

Go through the loss analysis with an existing charger EVM design
TI Information – Selective Disclosure
Linear Chargers
VIN
14.0
VBAT
3S Vbat=9.0V, Vin=13.2
ICHG
Adapter
Linear
Charger
Battery
Ploss (W)
12.0
+
2S Vbat=6.0, Vin=8.8V
10.0
8.0
6.0
4.0
PLOSS  VIN  VBAT   I chg

Simple and low cost

High loss
–

2.0
0.0
0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3
Ichg (A)
Difference of the adaptor and battery voltage
Only for small current
- The charging current is limited due to the high loss
TI Information – Selective Disclosure
Advantage of Switching Chargers
VIN
2.0
VBAT
3S Vbat=12.6V, Vin=19V
Battery
Adapter
Switching Charger
Ploss (W)
+
2S Vbat=8.4V, Vin=19V
1.5
1.0
0.5
0.0

–
–

0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3
High efficiency
Wide range input voltage
High output current
Ichg (A)
High output current
Need to understand the loss and optimize the efficiency
TI Information – Selective Disclosure
A Switching Charger and the Loss Components
RS1
Cin
Q1
L
Q2
Driver and
Controller
RSNS
+
Cout
Conduction
(IR)
Switching
Gate Driver
Other
Q1
√
√
√
Qrr Loss
Q2
√
√
Dead time
Loss
Inductor
√
Rs1, Rs2
√
Core Loss
IC
PCB
Gate Driver
√
TI Information – Selective Disclosure
Circuit under Study --- bq24715 NVDC-1 Charger
L
RS1 Q1
System
Cin
Cout
bq24715

Battery
Pack
Key features
–
–
–

Q2
Qbat
NVDC-1 Charger
Extreme low quiescent current to meet Energy Star Requirement
Ultra fast transient 100us to supplement mode to prevent adaptor
crash during turbo boost operation
Operation Condition
–
–
Vin=19V, Vo=8.4V, Io=6A
Fs=800KHz
TI Information – Selective Disclosure
How to Select MOSFET
TI Information – Selective Disclosure
MOSFET Losses
Conduction Loss
VDS
Switching Loss
Gate driver Loss

MOSFET is equivalent to a R when it is fully on

Loss is with I-V overlapping during the On-off transition

Capacitor charge and discharge
How to find the information on the DS
TI Information – Selective Disclosure
ID
VGS
Rdson Dependency on the Gate Drive Voltage
CSD17308Q3


When the switch is on, it is equivalent to a resistor RDS_on. Which
determines the conduction loss
RDS_on is a function of the driver voltage
TI Information – Selective Disclosure
Rdson Dependency on the Temp

RDS_on is a strong function of temperature. At 150oC junction temperature,
the temp coefficient is around 1.4 to 1.5

The conduction loss calculation must take the temperature coefficient into
consideration
TI Information – Selective Disclosure
Calculation the Conduction Loss
IQ1
L
IQ2
Q2
Q1
IOUT
∆IL
MOSFET Q1
IOUT
C
IQ1
D·T
Pcond  I
2
rms
R DS _ on Cotemp
1 DT 2
1 2
 2
I rms 
I o dt  D I OUT  I L 

0
T
12



The conduction loss for Q1 and Q2 can be calculated

It starts with a assumed temperature and iteration
TI Information – Selective Disclosure
T
Gate Charge and Switching loss
ID
VDS
Cgd
VGS
Rg
Cds
Cgs
VGS(th)
QGS2 QGD QGodr
QGS(th)



Qsw determines the switching loss
FOM = RDS_on x Qsw
The test condition is important
Qgs
TI Information – Selective Disclosure
Qsw
t
QGD is a Function of VDS
Increased VDS

Qgd is a function of VDS and Qg is a function of VGS

The comparison of the Qgd should be under the same Vds conditions

Some MOSFET venders specify Qgd at low Vds, resulting in better
data sheets, but not better performance
TI Information – Selective Disclosure
QGD a Function of VDS
?


The Rdson and Qgd are similar
The test conditions are different
TI Information – Selective Disclosure
Find the Correct QGD

Need to use the charge graph to
determine the charge under
certain conditions

The charge under the same test
condition is shown below (30%
higher Qgd)
53
13
TI Information – Selective Disclosure
15V
Switching Loss Accurate Formula

Lin
Switching loss calculation
assumes linear transition
Da
Psw _ Qgd
12V
1
  I D VDS  t1  Fs
2
0
10K

Ig
The voltage transitions are nonlinear,
which can be included in Kv:
T
Kv
Vds
(5V/div)
Vgs
(1V/div)

t1
5s/div
V (t )dt


 0 .5
0
d
T  Vin
Kv is about from 0.27 to 0.35 for
most of the devices
Psw _ Qgd  K vI o  Vin  t1  Fs
TI Information – Selective Disclosure
Gate Drive Loss
L
Q1
bq24715
Q2
C
Gate Driver
Pgate_ Q1  Qg _ Q1 VDRV _ Q1  f SW
Pdrive  Pgate_ Q1  Pgate_ Q 2
Gate driver loss is the energy of the gate charge dissipated on
the resistance of the driver loop
 Gate driver loss is proportional to the gate charge and switching
frequency

TI Information – Selective Disclosure
Body Diode Conduction Loss
Vgs_Q1
Q1
L
Q2
Vbd_Q2 C
t
Vgs_Q2
Vds_Q2
t
Vbd
t
Ibd_Q2
ID_Q2
t
tDT
PBD _ Q 2  VBD  I OUT  (t DT1  t DT 2 )  f SW


The typical dead time is 20-40ns
The dead time loss impact becomes significant at high
switching frequency
TI Information – Selective Disclosure
MOSFET Selection vs. Loss
Conduction
(IR)
Switching
Gate Driver
Other
Q1
0.21
0.49
0.06
0.14 (Qrr)
Q2
0.24
0.06
0.13 (DT)
Inductor
√
Rs1
√
Core Loss
IC
PCB
Gate Driver
√
The table above shows the loss breakdown
 The selection is a tradeoff of cost and performance
 The optimized design is to minimize the loss for given MOSFETS

TI Information – Selective Disclosure
How to Select Inductor
TI Information – Selective Disclosure
Inductance Selection

30% to 40% peak-to-peak current at the worst scenario
VIN  VBAT VBAT 1
L
, I ripple  30% ICHG
I ripple VIN f s

Selection Consideration
–
–
–

Ipeak < Inductor Isat
Low DCR
Size such as low profile
Use table in Datasheet to select
TI Information – Selective Disclosure
Inductor and the Loss
2525CZ 3.3uH (6.9mm x 6.5mm x 3mm)

Manufacturers provide calculation tools
Core loss calculation: http://www.vishay.com/docs/34252/ihlpse.pdf
Copper loss
Inductor
Switching
1.11
Gate Driver
Other
0.15 (core)
TI Information – Selective Disclosure
Sensing Resistors and IC Loss
TI Information – Selective Disclosure
Sensing Resistor

Selection Consideration
–
–
Accuracy : requiring high value of sensing resistance
The main source of the error is the offset of the comparator
bq24715
PARAMETER
INPUT CURRENT
REGULATION (0-125C)
TEST CONDITION
10mΩ current
sensing resistor
MIN
TYP
MAX
UNIT
3937
4096
4219
V
3
%
-3
–
Competition needs 20mΩ sensing resistor to achieve the
same accuracy
–
Power dissipation: requiring low value of sensing resistance
2
2
PRsens  I IN
R SENSE _ IN  I CHG
R SENSE _ CHG
TI Information – Selective Disclosure
bq24715 Quiescent Current Efficiency
~80mW
<500mW
<215mW
System
Q1 Q2
bq24715
Qbat
Batter Pack
bq24715
PARAMETER
TEST CONDITION
Standby Quiescent Vin=20V, Vbat=12.6V
Current
TJ = -20 to 85°C. No switching

MIN
TYP
MAX
UNIT
0.7
mA
Standby current
–
–
Crucial to the light load efficiency and meet the Energy Star
requirement
Competition has a maximum 5mA
TI Information – Selective Disclosure
Loss Breakdown
Conduction
(IR)
Switching
Gate Driver
Other
Q1
0.21
0.46
0.06
0.14 (Qrr)
Q2
0.24
0.06
0.23 (DT)
Rs1
0.09
Inductor
1.11
0.15 (Core)
0.12
IC
0.013 (Bias)
0.1
PCB
28%

Q1
3%
Q2
5% IC
16%
The loss has a good match
–
–
3% Rsen
–
L
The calculated loss is 2.86W
The measured loss is about
2.98W
Can be verified at different
operation points
45%
TI Information – Selective Disclosure
Summary

MOSFET selection is based on the loss optimization and cost
trade off. The loss modeling of a MOSFET is analyzed:
1.
Conduction loss
2.
Switching loss
3.
Dead time loss
4.
Gate drive

The selection of a Inductor and the tradeoff is discussed

Other loss in a charger circuit breakdown and the impact are
addressed

The EVM loss breakdown is conducted
TI Information – Selective Disclosure