ON NCP5218MNR2G 2−in−1 notebook ddr power controller Datasheet

NCP5218
2−in−1 Notebook DDR
Power Controller
The NCP5218 2−in−1 Notebook DDR Power Controller is
specifically designed as a total power solution for notebook DDR
memory system. This IC combines the efficiency of a PWM
controller for the VDDQ supply with the simplicity of linear
regulators for the VTT termination voltage and the buffered low noise
reference. This IC contains a synchronous PWM buck controller for
driving two external NFETs to form the DDR memory supply
voltage (VDDQ). The DDR memory termination regulator output
voltage (VTT) and the buffered VREF are internally set to track at the
half of VDDQ. An internal power good voltage monitor tracks VDDQ
output and notifies the user whether the VDDQ output is within target
range. Protective features include soft−start circuitries, undervoltage
monitoring of supply voltage, VDDQ overcurrent protection, VDDQ
overvoltage and undervoltage protections, and thermal shutdown.
The IC is packaged in DFN22.
Features
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Incorporates VDDQ, VTT Regulator, Buffered VREF
Adjustable VDDQ Output
VTT and VREF Track VDDQ/2
Operates from Single 5.0 V Supply
Supports VDDQ Conversion Rails from 4.5 V to 24 V
Power−saving Mode for High Efficiency at Light Load
Integrated Power FETs with VTT Regulator Sourcing/Sinking
1.5 A DC and 2.4 A Peak Current
Requires Only 20 mF Ceramic Output Capacitor for VTT
Buffered Low Noise 15 mA VREF Output
All External Power MOSFETs are N−channel
<5.0 mA Current Consumption During Shutdown
Fixed Switching Frequency of 400 kHz
Soft−start Protection for VDDQ and VTT
Undervoltage Monitor of Supply Voltage
Overvoltage Protection and Undervoltage Protection for VDDQ
Short−circuit Protection for VDDQ and VTT
Thermal Shutdown
Housed in DFN22
This is a Pb−Free Device
Typical Applications
• Notebook DDR/DDR2 Memory Supply and Termination Voltage
• Active Termination Busses (SSTL−18, SSTL−2, SSTL−3)
© Semiconductor Components Industries, LLC, 2007
January, 2007 − Rev. 0
1
http://onsemi.com
MARKING
DIAGRAM
22
DFN22
MN SUFFIX
CASE 506AF
1
NCP5218
AWLYYWW
G
1
NCP5218= Specific Device Code
A
= Assembly Location
WL
= Wafer Lot
YY
= Year
WW
= Work Week
G
= Pb−Free Package
PIN CONNECTIONS
VDDQEN
VTTEN
FPWM
SS
VTTGND
VTT
VTTI
FBVTT
AGND
DDQREF
VCCA
(Top View)
PGND
BGDDQ
VCCP
SWDDQ
TGDDQ
BOOST
OCDDQ
PGOOD
VTTREF
FBDDQ
COMP
NOTE: Pin 23 is the thermal pad on
the bottom of the device.
ORDERING INFORMATION
Device
Package
Shipping†
NCP5218MNR2G
DFN22
(Pb−Free)
2500 Tape & Reel
†For information on tape and reel specifications,
including part orientation and tape sizes, please
refer to our Tape and Reel Packaging Specification
Brochure, BRD8011/D.
Publication Order Number:
NCP5218/D
NCP5218
VDDQEN
VDDQEN
RL1
VTTEN
VTTEN
OCDDQ
FPWM
FPWM
5VCC
SS
BOOST
CSS
PWRGD
0.9 V, 1.5 A
VTT
VIN
VCCP
5VCC
M1
PGOOD
TGDDQ
L
VTT
COUT2
Ceramic
10 mF x2
FBVTT
4.5 V to 24 V
(Battery/
Adapter)
VDDQ
1.8 V, 10 A
SWDDQ
NCP5218
M2
1.8 mH
COUT1
POSCAP
150 mF x2
BGDDQ
PGND1
VTTGND
5VCC
COMP
VCCA
CZ1
RZ1
VREF
0.9 V, 15 mA
CP1
VTTREF
CZ2
R1
RZ2
FBDDQ
R2
DDQREF
AGND
VTTI
Figure 1. Typical Application Diagram
http://onsemi.com
2
NCP5218
5VCC
VIN
VREF
VREFGD
THERMAL
SHUTDOWN
TSD
VDDQEN
VTTEN
VTTEN
FPWM
FPWM
VOCDDQGD
VCCA
5VCC
VCCP
CONTROL
LOGIC
VCCAGD
VCCA
CBULK
VCCP
VBOOST
BOOST
FAULT
+
−
VREF
INREGDDQ
ILIM
RL1
+
−
IREF
OCDDQ
VBOOST
VOCDDQ
VREF
CDCPL
VDDQEN
CBOOST
VOLTAGE &
CURRENT
REFERENCE
VDDQ
PWM
LOGIC
+
−
M3
TGDDQ
FBDDQ
L
SWDDQ
VDDQEN
VCCA
VTTEN
SS
Power−
Saving
Loop
Control
SWDDQ
VCCP
NEGATIVE CURRENT
DETECTION
5VCC
BGDDQ
PGND
PGND
VREF
PGOOD
OSC
+ −
+
−
UVLO
VFBDDQ
OVLO
VFBDDQ
+
−
VREF
COMP
PGND
VREF
VOCDDQ
Adaptive
Ramp
A
COUT1
M4
+
−
PWM−
COMP
CZ1
CZ2
CP1
RZ1
+
−
RZ2
R1
FBDDQ
R2
+
−
SC2PWR
VTTREF
Deadband
Control
+
−
DDQREF
Current
Limit &
Soft−Start
VTTI
VCCA
M1
VTTI
VDDQEN
VTTEN
INREGDDQ
VTT
VTT
Regulation
Control
VTTGND
VTT
VCCA
COUT2
M2
SC2GND
PGND
+
−
VTTREF
COUT3
VDDQ
VTTGND
VTTGND
VTTGND
FBVTT
GND
AGND
VTTGND
Figure 2. Detailed Block Diagram
http://onsemi.com
3
NCP5218
PIN FUNCTION DESCRIPTION
Pin
Symbol
Description
1
VDDQEN
2
VTTEN
VTT regulator enable input. High to enable.
3
FPWM
Forced PWM enable input. Low to enable forced PWM mode and disable power−saving mode.
VDDQ regulator enable input. High to enable.
4
SS
5
VTTGND
VDDQ Soft−start capacitor connection to ground.
6
VTT
VTT regulator output.
7
VTTI
Power input for VTT regulator which is normally connected to the VDDQ output of the buck regulator.
8
FBVTT
VTT regulator feedback pin for closed loop regulation.
9
AGND
Analog ground connection and remote ground sense.
10
DDQREF
11
VCCA
5.0 V supply input for the IC’s control and logic section, which is monitored by undervoltage lock out
circuitry.
12
COMP
VDDQ error amplifier compensation node.
13
FBDDQ
VDDQ regulator feedback pin for closed loop regulation.
14
VTTREF
DDR reference voltage output.
15
PGOOD
Power good signal open−drain output.
16
OCDDQ
Overcurrent sense and program input for the high−side FET of VDDQ regulator. Also the battery voltage
input for the internal ramp generator to implement the voltage feedforward rejection to the input voltage
variation. This pin must be connected to the VIN through a resistor to perform the current limit and
voltage feedforward functions.
17
BOOST
Positive supply input for high−side gate driver of VDDQ regulator and boost capacitor connection.
18
TGDDQ
Gate driver output for VDDQ regulator high−side N−Channel power FET.
19
SWDDQ
VDDQ regulator inductor driven node, return for high−side gate driver, and current limit sense input.
20
VCCP
Power supply for the VDDQ regulator low−side gate driver and also supply voltage for the bootstrap
capacitor of the VDDQ regulator high−side gate driver supply.
21
BGDDQ
Power ground for the VTT regulator.
External reference input which is used to regulate VTT and VTTREF to 1/2VDDQREF.
Gate driver output for VDDQ regulator low−side N−Channel power FET.
22
PGND
Power ground for the VDDQ regulator.
23
THPAD
Copper pad on bottom of IC used for heatsinking. This pin should be connected to the ground plane
under the IC.
http://onsemi.com
4
NCP5218
MAXIMUM RATINGS
Rating
Symbol
Value
Unit
VCCA, VCCP
−0.3, 6.0
V
VBOOST−VSWDDQ
−0.3, 6.0
V
VIO
−0.3, 6.0
V
Overcurrent Sense Input (Pin 16) to AGND (Pin 9)
VOCDDQ
27
V
Switch Node (Pin 19)
VSWDDQ
−4.0 (<100 ns),
−0.3 (dc), 32
V
VTGDDQ −
VSWDDQ
−2.0 (< 100 ns)
−0.3 (dc), 6.0
V
VBGDDQ
−2.0 (< 100 ns)
−0.3 (dc), 6.0
V
PGND (Pin 22), VTTGND (Pin 5) to AGND (Pin 9)
VGND
−0.3, 0.3
V
Thermal Characteristics
DFN22 Plastic Package
Thermal Resistance, Junction−to−Ambient
RqJA
35
_C/W
Power Supply Voltage (Pin 11, 20) to AGND (Pin 9)
High−Side Gate Drive Supply: BOOST (Pin 17) to SWDDQ (Pin 19)
Input/Output Pins to AGND (Pin 9)
Pins 1−4, 6−8, 10, 12−15
High−Side FET Gate Drive Voltage: TGDDQ (Pin 18) To SWDDQ (Pin 19)
Low−Side FET Gate Drive Voltage: BGDDQ (Pin 21) To PGND (Pin 22)
Operating Junction Temperature Range
TJ
0 to +150
_C
Operating Ambient Temperature Range
TA
−40 to +85
_C
Storage Temperature Range
Tstg
−55 to +150
_C
Moisture Sensitivity Level
MSL
1
−
Stresses exceeding Maximum Ratings may damage the device. Maximum Ratings are stress ratings only. Functional operation above the
Recommended Operating Conditions is not implied. Extended exposure to stresses above the Recommended Operating Conditions may affect
device reliability.
1. This device series contains ESD protection and exceeds the following tests:
Human Body Model (HBM) ≤2.0 kV per JEDEC standard: JESD22–A114 except Pin 17 which is ≤ 1 kV.
Machine Model (MM) ≤200 V per JEDEC standard: JESD22–A115 except Pin 17 which is ≤ 150 V.
2. Latchup Current Maximum Rating: ≤150 mA per JEDEC standard: JESD78.
3. Pin 16 (OCDDQ) must be pulled high to VIN through a resistor.
http://onsemi.com
5
NCP5218
ELECTRICAL CHARACTERISTICS (VIN = 12 V, TA = −40 to 85_C, VCCA = VCCP = VBOOST − VSWDDQ = 5.0 V, L = 1.8 mH,
COUT1 = 150 mF x 2, COUT2 = 22 mF x 2, RL1 = 5.6 kW, R1 = 4.3 kW, R2 = 3.3 kW, RZ1 = 10 kW, RZ2 = 130 W, CP1 = 100 pF, CZ1 = 2.2 nF,
CZ2 = 4.7 nF, for min/max values unless otherwise noted. Typical values are at TA = 25_C.)
Characteristic
Symbol
Test Conditions
Min
Typ
Max
Unit
VIN
−
4.5
−
24
V
VCCA Operating Voltage
VCCA
−
4.5
5.0
5.5
V
VCCP Operating Voltage
VCCP
−
4.5
5.0
5.5
V
VCCA Quiescent Supply Current in S0
IVCCA_S0
VDDQEN = 5.0 V, VTTEN = 5.0 V
−
3.5
10
mA
VCCA Quiescent Supply Current in S3
IVCCA_S3
VDDQEN = 5.0 V, VTTEN = 0 V
−
0.9
5.0
mA
VCCA Shutdown Current
IVCCA_SD
VDDQEN = 0 V, VTTEN = 0 V,
TA = 25°C
−
1.0
4.0
mA
VCCP Quiescent Supply Current in S0
IVCCP_S0
VDDQEN = 5.0 V, VTTEN = 5.0 V,
TGDDQ and BGDDQ Open
−
−
20
mA
VCCP Quiescent Supply Current in S3
IVCCP_S3
VDDQEN = 5.0 V, VTTEN = 0 V,
TGDDQ and BGDDQ Open
−
−
20
mA
VCCP Shutdown Current
IVCCP_SD
VDDQEN = 0 V, VTTEN = 0 V
−
1.0
2.0
mA
VCCAUV−
Falling Edge
−
3.7
4.1
V
SUPPLY VOLTAGE
Input Voltage
SUPPLY CURRENT
UNDERVOLTAGE MONITOR
VCCA UVLO Lower Threshold
VCCA UVLO Hysteresis
VCCAUVHYS
−
−
0.35
−
V
VOCDDQ UVLO Upper Threshold
VOCDDQUV+
Rising Edge
−
3.0
4.4
V
VOCDDQUVHYS
−
−
0.4
−
V
VOCDDQ UVLO Hysteresis
THERMAL SHUTDOWN
Thermal Trip Point
TSD
(Note 4)
−
150
−
_C
TSDHYS
(Note 4)
−
25
−
_C
VFBDDQ
TA = 25°C
TA = −40 to 85°C
0.788
0.784
0.8
0.8
0.812
0.816
V
Ifb
VFBDDQ = 0.8 V
−
−
1.0
mA
Oscillator Frequency
FSW
−
340
400
460
kHz
Ramp Amplitude Voltage
Vramp
VIN = 5.0 V (Note 4)
−
1.25
−
V
dVRAMP/dVIN
−
−
45
−
mV/V
IOC
VOCDDQ = 4.0 V
26
31
36
mA
OCDDQ Pin Current Sink
Temperature Coefficient
TCIOC
TA = −40 to 85°C
−
3200
−
ppm/
_C
Minimum On Time
tonmin
−
−
150
−
ns
Maximum Duty Cycle
Dmax
VIN = 5.0 V
VIN = 15 V
VIN = 24 V
−
−
−
90
50
32
−
−
−
%
Iss
VDDQEN = 5.0 V, Vss = 0 V
2.8
4.0
5.2
mA
Overvoltage Trip Threshold
FBOVPth
With Respect to Error
Comparator Threshold of 0.8 V
115
130
−
%
Undervoltage Trip Threshold
FBUVPth
With Respect to Error
Comparator Threshold of 0.8 V
−
65
75
%
Hysteresis
VDDQ SWITCHING REGULATOR
FBDDQ Feedback Voltage, Control Loop in
Regulation
Feedback Input Current
Ramp Amplitude to VIN Ratio
OCDDQ Pin Current Sink
Soft−Start Current
4. Guaranteed by design, not tested in production.
http://onsemi.com
6
NCP5218
ELECTRICAL CHARACTERISTICS (continued) (VIN = 12 V, TA = −40 to 85_C, VCCA = VCCP = VBOOST − VSWDDQ = 5.0 V,
L = 1.8 mH, COUT1 = 150 mF x 2, COUT2 = 22 mF x 2, RL1 = 5.6 kW, R1 = 4.3 kW, R2 = 3.3 kW, RZ1 = 10 kW, RZ2 = 130 W, CP1 = 100 pF,
CZ1 = 2.2 nF, CZ2 = 4.7 nF, for min/max values unless otherwise noted. Typical values are at TA = 25_C.)
Characteristic
Symbol
Test Conditions
Min
Typ
Max
Unit
GAIN
(Note 5)
−
70
−
dB
Unity Gain Bandwidth
Ft
COMP_GND = 220 nF,
1.0 W in Series (Note 5)
−
2.0
−
MHz
Slew Rate
SR
(Note 5)
−
3.0
−
V/mS
TGDDQ Gate Pull−HIGH Resistance
RH_TG
VBOOST − VSWDDQ = 5.0 V,
VTGDDQ − VSWDDQ = 4.0 V
−
1.8
4.0
W
TGDDQ Gate Pull−LOW Resistance
RL_TG
VBOOST − VSWDDQ = 5.0 V,
VTGDDQ − VSWDDQ = 1.0 V
−
1.8
4.0
W
BGDDQ Gate Pull−HIGH Resistance
RH_BG
VCCP = 5.0 V, VBGDDQ = 4.0 V
−
1.8
4.0
W
BGDDQ Gate Pull−LOW Resistance
RL_BG
VCCP = 5.0 V, VBGDDQ = 1.0 V
−
0.9
3.0
W
dVTT0
1/2VDDQREF – VTT,
VDDQREF = 2.5 V,
IVTT = 0 to 2.4 A
(Sink Current)
IVTT = 0 to –2.4 A
(Source Current)
ERROR AMPLIFIER
DC Gain
GATE DRIVERS
VTT ACTIVE TERMINATOR
VTT with Respect to 1/2VDDQREF
1/2VDDQREF – VTT,
VDDQREF = 1.8 V,
IVTT = 0 to 2.0 A
(Sink Current)
IVTT = 0 to –2.0 A
(Source Current)
DDQREF Input Resistance
mV
−30
−
−
−
−
30
mV
−30
−
−
−
−
30
DDQREF_R
VDDQREF = 2.5 V
40
55
75
kW
Source Current Limit
ILIMVTsrc
−
2.5
3.0
−
A
Sink Current Limit
ILIMVTsnk
−
2.5
3.0
−
A
Soft−Start Source Current Limit
ILIMVTSS
−
−
1.0
−
A
Maximum Soft−Start Time
tssvttmax
−
−
0.32
−
ms
VTTREF Source Current
IVTTR
VDDQREF = 1.8 V or 2.5 V
15
−
−
mA
VTTREF Accuracy Referred to 1/2VDDQREF
dVTTR
1/2VDDQREF – VTTR,
VDDQREF = 2.5 V,
IVTTR = 0 mA to 15 mA
−25
−
25
mV
1/2VDDQREF – VTTR,
VDDQREF = 1.8 V,
IVTTR = 0 mA to 15 mA
−18
−
18
mV
VTTREF OUTPUT
5. Guaranteed by design, not tested in production.
http://onsemi.com
7
NCP5218
ELECTRICAL CHARACTERISTICS (continued) (VIN = 12 V, TA = −40 to 85_C, VCCA = VCCP = VBOOST − VSWDDQ = 5.0 V,
L = 1.8 mH, COUT1 = 150 mF x 2, COUT2 = 22 mF x 2, RL1 = 5.6 kW, R1 = 4.3 kW, R2 = 3.3 kW, RZ1 = 10 kW, RZ2 = 130 W, CP1 = 100 pF,
CZ1 = 2.2 nF, CZ2 = 4.7 nF, for min/max values unless otherwise noted. Typical values are at TA = 25_C.)
Characteristic
Symbol
Test Conditions
Min
Typ
Max
Unit
VDDQEN Pin Threshold High
VDDQEN_H
−
1.4
−
−
V
VDDQEN Pin Threshold Low
VDDQEN_L
−
−
−
0.5
V
VDDQEN Pin Input Current
IIN_
VDDQEN
VDDQEN = 5.0 V
−
−
1.0
mA
VTTEN Pin Threshold High
VTTEN_H
−
1.4
−
−
V
VTTEN Pin Threshold Low
VTTEN_L
−
−
−
0.5
V
VTTEN Pin Input Current
IIN_VTTEN
VDDQEN = VTTEN = 5.0 V
−
−
1.0
mA
FPWM Pin Threshold High
FPWM_H
−
1.4
−
−
V
FPWM Pin Threshold Low
FPWM_L
−
−
−
0.5
V
FPWM Pin Input Current
IIN_FPWM
VDDQEN = VTTEN =FPWM
= 5.0 V
−
−
1.0
mA
PGOOD Pin ON Resistance
PGOOD_R
I_PGOOD = 5.0 mA
−
70
−
W
PGOOD Pin OFF Current
PGOOD_LK
−
−
−
1.0
mA
thold
(Note 6)
−
−
200
ms
CONTROL SECTION
PGOOD LOW−to−HIGH Hold Time, for S5 to S0
6. Guaranteed by design, not tested in production.
http://onsemi.com
8
NCP5218
IVCCA_S3, QUIESCENT CURRENT IN S3 (mA)
IVCCA_S0, QUIESCENT CURRENT IN S0 (mA)
TYPICAL OPERATING CHARACTERISTICS
4.0
3.8
3.6
3.4
3.2
3.0
−40
−15
10
35
60
TA, AMBIENT TEMPERATURE (°C)
85
1.0
0.8
0.6
0.4
0.2
0.0
−40
10
8
6
4
2
0
−40
−15
10
35
60
TA, AMBIENT TEMPERATURE (°C)
85
450
425
400
375
350
−40
−15
10
35
60
TA, AMBIENT TEMPERATURE (°C)
85
Figure 6. Switching Frequency in S0
vs. Ambient Temperature
5.0
0.90
ISS, SOFT−START CURRENT (mA)
VFBDDQ, VDDQ FEEDBACK VOLTAGE (V)
Figure 5. VCCA Shutdown Current
vs. Ambient Temperature
0.85
0.80
0.75
0.70
−40
85
Figure 4. VCCA Quiescent Current in S3
vs. Ambient Temperature
FSW, SWITCHING FREQUENCY IN S0 (kHz)
IVCCA_SD, SHUTDOWN CURRENT (mA)
Figure 3. VCCA Quiescent Current in S0
vs. Ambient Temperature
−15
10
35
60
TA, AMBIENT TEMPERATURE (°C)
−15
10
35
60
TA, AMBIENT TEMPERATURE (°C)
4.5
4.0
3.5
3.0
−40
85
Figure 7. VDDQ Feedback Voltage
vs. Ambient Temperature
−15
10
35
60
TA, AMBIENT TEMPERATURE (°C)
Figure 8. Soft−Start Current
vs. Ambient Temperature
http://onsemi.com
9
85
NCP5218
TYPICAL OPERATING CHARACTERISTICS
1.820
VDDQ, VDDQ OUTPUT VOLTAGE (V)
VDDQ, VDDQ OUTPUT VOLTAGE (V)
1.810
1.815
1.810
1.805
1.805
IVDDQ = 100 mA
1.800
1.795
IVDDQ = 10 A
1.790
0
5
10
15
VIN, INPUT VOLTAGE (V)
20
VIN = 5 V
VIN = 24 V
1.795
VDDQ = 1.8 V
S0 Mode
TA = 25°C
1.785
1.780
1.800
VDDQ = 1.8 V
TA = 25°C
1.790
25
0
2
4
6
8
IVDDQ, VDDQ OUTPUT CURRENT (A)
Figure 10. VDDQ Output Voltage
vs. VDDQ Output Current
1.29
0.94
1.28
0.93
VTT, VTT OUTPUT VOLTAGE (V)
VTT, VTT OUTPUT VOLTAGE (V)
Figure 9. VDDQ Output Voltage
vs. Input Voltage
1.27
1.26
1.25
1.24
VIN = 5 V
1.23
1.22
1.21
−3.0
VDDQ = 2.5 V
TA = 25°C
−2.0
VIN = 24 V
−1.0
0.0
1.0
2.0
IVTT, VTT OUTPUT CURRENT (A)
0.92
0.91
0.90
0.89
VIN = 5 V
0.88
VIN = 24 V
0.87
0.86
−2.0
3.0
Figure 11. VTT Output Voltage (DDR)
vs. VTT Output Current
VDDQ = 1.8 V
TA = 25°C
−1.5
−1.0 −0.5 0.0
0.5
1.0
IVTT, VTT OUTPUT CURRENT (A)
1.5
2.0
Figure 12. VTT Output Voltage (DDR2)
vs. VTT Output Current
VTTR, VTTR OUTPUT VOLTAGE (V)
0.910
VTTR, VTTR OUTPUT VOLTAGE (V)
1.260
1.255
0.905
1.250
0.900
VIN = 5 V
1.245
1.240
10
VDDQ = 2.5 V
TA = 25°C
0
VIN = 24 V
5
10
IVTTR, VTTR OUTPUT CURRENT (mA)
VIN = 5 V
0.895
0.890
15
VDDQ = 1.8 V
TA = 25°C
0
Figure 13. VTTR Output Voltage (DDR)
vs. VTTR Output Current
VIN = 24 V
5
10
IVTTR, VTTR OUTPUT CURRENT (mA)
Figure 14. VTTR Output Voltage (DDR2)
vs. VTTR Output Current
http://onsemi.com
10
15
NCP5218
TYPICAL OPERATING CHARACTERISTICS
100
VIN = 5 V
VIN = 12 V
VIN = 20 V
90
80
EFFICIENCY OF VDDQ (%)
EFFICIENCY OF VDDQ (%)
100
with power−saving
without power−saving
70
60
VDDQ = 2.5 V
Freq = 400 kHz max
TA = 25°C
50
0.1
1.0
10
IVDDQ, VDDQ OUTPUT CURRENT (A)
100
VIN = 5 V
VIN = 12 V
VIN = 20 V
90
80
with power−saving
without power−saving
70
60
VDDQ = 1.8 V
Freq = 400 kHz max
TA = 25°C
50
0.1
1.0
10
IVDDQ, VDDQ OUTPUT CURRENT (A)
Figure 16. VDDQ Efficiency (DDR2)
vs. VDDQ Output Current
Figure 15. VDDQ Efficiency (DDR)
vs. VDDQ Output Current
VIN
20V/div
VIN
20V/div
VDDQ
1V/div
1V/div
VDDQ
VTT
1V/div
VTT
1V/div
VTTR
1V/div
VTTR
1V/div
VDDQEN = High; VTTEN = High; VIN =20 V to 0 V
VDDQEN = High; VTTEN = High; VIN = 0 V to 20 V
Figure 17. Powerup Waveforms
VDDQEN
Figure 18. Powerdown Waveforms
5V/div
VDDQEN
5V/div
1V/div
VDDQ
VTTR
1V/div
VTTR
1V/div
PGOOD
5V/div
PGOOD
5V/div
1V/div
VDDQ
VDDQEN = 0 V to 5 V
VDDQEN = 5 V to 0 V
Figure 19. VDDQ, VTTR Startup Waveforms
Figure 20. VDDQ, VTTR Shutdown Waveforms
http://onsemi.com
11
100
NCP5218
TYPICAL OPERATING CHARACTERISTICS
VTTEN
5V/div
VTTEN
5V/div
VTT
1V/div
VTT
1V/div
500mA/div
IVTTI
IVTTI
VDDQEN = High; VTT Loaded with 4.7 W to GND
VDDQEN = High; VTT Loaded with 4.7 W to GND
Figure 21. VTT Startup Waveforms
Figure 22. VTT Shutdown Waveforms
100mV/div
VDDQ
500mA/div
VTT
100mV/div
VDDQ
1V/div
1V/div
VTT
VTTR
VTTEN
50mV/div
VTTR
50mV/div
5V/div
FPWM
5V/div
IVDDQ = 50 mA, IVTT = 100 mA, IVTTR = 5 mA
IVDDQ = 50mA, IVTT = 100mA, IVTTR = 5mA, VTTEN = 0V
Figure 23. S0−S3−S0 Transition Waveforms
Figure 24. PS−FPWM−PS Transition
Waveforms
100mV/div
VDDQ
VDDQ
100mV/div
VTT
50mV/div
VTT
50mV/div
VTTR
50mV/div
VTTR
50mV/div
5A/div
IVDDQ
5A/div
IVDDQ
IVDDQ = 0 A−7 A, IVTT = 1.5 A, IVTTR = 15 mA
IVDDQ = 7 A−0 A, IVTT = 1.5 A, IVTTR = 15 mA
Figure 25. VDDQ Load Transient
Figure 26. VDDQ Load Transient
http://onsemi.com
12
NCP5218
TYPICAL OPERATING CHARACTERISTICS
VDDQ
100mV/div
VDDQ
100mV/div
VTT
50mV/div
VTT
50mV/div
VTTR
50mV/div
VTTR
50mV/div
IVTT
2A/div
IVTT
IVDDQ = 8 A, IVTT = 0 A to −2 A to 0 A, IVTTR = 15 mA
IVDDQ = 8 A, IVTT = 0 A to 2 A to 0 A, IVTTR = 15 mA
Figure 27. VTT Source Current Transient
VDDQ
VTT
VTTR
VIN
Figure 28. VTT Sink Current Transient
VDDQ
100mV/div
50mV/div
50mV/div
10V/div
VTT
50mV/div
VTTR
50mV/div
10V/div
IVDDQ = 0 A, IVTT = 0 A, IVTTR = 0 mA, VIN = 20 V to 7 V
Figure 29. Line Transient 7 V to 20 V at No Load
100mV/div
100mV/div
VTT
VIN
IVDDQ = 0 A, IVTT = 0 A, IVTTR = 0 mA, VIN = 7 V to 20 V
VDDQ
2A/div
Figure 30. Line Transient 20 V to 7 V at No Load
VDDQ
100mV/div
50mV/div
VTT
50mV/div
VTTR
50mV/div
VTTR
50mV/div
VIN
10V/div
VIN
10V/div
IVDDQ = 10A, IVTT = 1.5A, IVTTR = 15mA, VIN = 20V to 7V
IVDDQ = 10A, IVTT = 1.5A, IVTTR = 15mA, VIN = 7V to 20V
Figure 31. Line Transient 7 V to 20 V at Full Load
Figure 32. Line Transient 20 V to 7 V at Full Load
http://onsemi.com
13
NCP5218
TYPICAL OPERATING CHARACTERISTICS
100mV/div
VDDQ
100mV/div
VTT
1V/div
VTT
1V/div
VTTR
50mV/div
VTTR
50mV/div
IVTT
5A/div
VDDQ
IVTT
5A/div
IVDDQ = 8 A, VTT shorts to VDDQ, IVTTR = 15 mA
IVDDQ = 8 A, VTT shorts to ground, IVTTR = 15 mA
Figure 33. VTT Short Circuit to Ground and
Recovery
Figure 34. VTT Short Circuit to VDDQ and
Recovery
VDDQ, 1V/div
VDDQ, 1V/div
VSWDDQ, 10V/div
VSWDDQ, 10V/div
VIN, 20V/div
VIN, 20V/div
IL, 10A/div
IL, 10A/div
Figure 35. VDDQ OCP by Short Circuit to
Ground
Figure 36. VDDQ OCP by Steady IVDDQ
Increase
VDDQ, 1V/div
VSWDDQ, 10V/div
VIN, 20V/div
IL, 10A/div
Figure 37. VDDQ OCP by Start into a Short
Circuit
http://onsemi.com
14
NCP5218
DETAILED OPERATING DESCRIPTION
General
VDDQ output voltage is divided down and fed back to the
inverting input of an internal error amplifier through
FBDDQ pin to close the loop at VDDQ = VFBDDQ ×
(1 + R1/R2). This amplifier compares the feedback voltage
with an internal VREF (= 0.800 V) to generate an error
signal for the PWM comparator. This error signal is further
compared with a fixed frequency RAMP waveform
derived from the internal oscillator to generate a
pulse−width−modulated signal. This PWM signal drives
the external N−Channel Power FETs via the TGDDQ and
BGDDQ pins. External inductor L and capacitor COUT1
filter the output waveform. The VDDQ output voltage
ramps up at a pre−defined soft−start rate when the IC enters
state S0 from S5. When in normal mode, and regulation of
VDDQ is detected, signal INREGDDQ will go HIGH to notify
the control logic block.
Input voltage feedforward is implemented to the RAMP
signal generation to reject the effect of wide input voltage
variation. With input voltage feedforward, the amplitude of
the RAMP is proportional to the input voltage.
For enhanced efficiency, an active synchronous switch is
used to eliminate the conduction loss contributed by the
forward voltage of a diode or Schottky diode rectifier.
Adaptive nonoverlap timing control of the complementary
gate drive output signals is provided to reduce large
shoot−through current that degrades efficiency.
The NCP5218 2−in−1 Notebook DDR Power Controller
combines the efficiency of a PWM controller for the VDDQ
supply, with the simplicity of using a linear regulator for the
VTT termination voltage power supply. The VDDQ output
can be adjusted through the external potential divider,
while the VTT is internally set to track half VDDQ.
The inclusion of VDDQ power good voltage monitor,
soft−start, VDDQ overcurrent protection, VDDQ
overvoltage and undervoltage protections, supply
undervoltage monitor, and thermal shutdown makes this
device a total power solution for high current DDR memory
system. The IC is packaged in DFN22.
Control Logic
The internal control logic is powered by VCCA. The IC
is enabled whenever VDDQEN is high (exceed 1.4 V). An
internal bandgap voltage, VREF, is then generated. Once
VREF reaches its regulation voltage, an internal signal
VREFGD will be asserted. This transition wakes up the
supply undervoltage monitor blocks, which will assert
VCCAGD if VCCA voltage is within certain preset levels.
The control logic accepts external signals at VCCA,
OCDDQ, VDDQEN, VTTEN, and FPWM pins to control
the operating state of the VDDQ and VTT regulators in
accordance with Table 1. A timing diagram is shown in
Figure 38.
Tolerance of VDDQ
VDDQ Switching Regulator in Normal Mode (S0)
The tolerance of VFBDDQ and the ratio of external
resistor divider R1/R2 both impact the precision of VDDQ.
With the control loop in regulation, VDDQ = VFBDDQ × (1
+ R1/R2). With a worst case (for all valid operating
conditions) VFBDDQ tolerance of "1.5%, a worst case
range of "2.5% for VDDQ = 1.8 V will be assured if the
ratio R1/R2 is specified as 1.2500 "1%.
The VDDQ regulator is a switching synchronous
rectification buck controller directly driving two external
N−Channel power FETs. An external resistor divider sets
the nominal output voltage. The control architecture is
voltage mode fixed frequency PWM with external
compensation and with switching frequency fixed at
400 kHz " 15%. As can be observed from Figure 1, the
Table 1. State, Operation, Input and Output Condition Table
Input Conditions
Operating Conditions
Output Conditions
Mode
VCCA
VOCDDQ
VDDQEN
VTTEN
FPWM
VDDQ
VTTREF
VTT
TGDDQ
BGDDQ
PGOOD
S5
Low
X
X
X
X
H−Z
H−Z
H−Z
Low
Low
Low
S5
X
Low
X
X
X
H−Z
H−Z
H−Z
Low
Low
Low
S0
High
High
High
High
X
Normal
Normal
Normal
Normal
Normal
H−Z
S3
High
High
High
Low
High
Standby
Normal
H−Z
Standby
(Power−
saving)
Standby
(Power−
saving)
H−Z
S3
High
High
High
Low
Low
Normal
Normal
H−Z
Normal
Normal
H−Z
S5
X
X
Low
X
X
H−Z
H−Z
H−Z
Low
Low
Low
VDDQ Regulator in Standby Mode (S3)
the switching frequency is reduced with the VDDQ output
current and the low−side FET is turned off after the
detection of negative inductor current, so as to enhance the
During state S3, a power−saving mode is activated when
the FPWM pin is pulled to VCCA. In power−saving mode,
http://onsemi.com
15
NCP5218
VTT Active Terminator in Standby Mode (S3)
efficiency of the VDDQ regulator at light loads. The
switching frequency can be reduced smoothly until it
reaches the minimum frequency at about 15 kHz.
Therefore, perceptible audible noise can be avoided at light
load condition.
In power−saving mode, the low−side MOSFET is turned
off after the detection of negative inductor current and the
converter cannot sink current. The power−saving mode can
be disabled by pulling the FPWM pin to ground. Then, the
converter operates in forced−PWM mode with fixed
switching frequency and ability to sink current.
VTT output is high−impedance in S3 mode.
Fault Protection of VTT Active Terminator
To provide protection for the internal FETs, bidirectional
current limit is implemented, preset at the minimum of
2.5 A magnitude.
Thermal Consideration of VTT Active Terminator
The VTT terminator is designed to handle large transient
output currents. If large currents are required for very long
duration, then care should be taken to ensure the maximum
junction temperature is not exceeded. The 5x6 DFN22 has
a thermal resistance of 35°C/W (dependent on air flow,
grade of copper, and number of vias). In order to take full
advantage from this thermal capability of this package, the
thermal pad underneath must be soldered directly onto a
PCB metal substrate to allow good thermal contact. It is
recommended that PCB with 2 oz. copper foil is used and
there should have 6 to 8 vias with 0.6 mm hole size
underneath the package’s thermal pad connecting the top
layer metal to the bottom layer metal and the internal layer
metal substrates of the PCB.
Fault Protection of VDDQ Regulator
During state S0 and S3, external resistor (RL1) between
OCDDQ and VIN sets the overcurrent trip threshold for the
high−side switch. An internal 31 mA current sink (IOC) at
OCDDQ pin establishes a voltage drop across this resistor
and develops a voltage at the non−inverting input of the
current limit comparator. The voltage at the non−inverting
input is compared to the voltage at SWDDQ pin when the
high−side gate drive is high after a fixed period of blanking
time (150 ns) to avoid false current limit triggering. When
the voltage at SWDDQ is lower than that at the
non−inverting input for 4 consecutive internal clock
cycles, an overcurrent condition occurs, during which, all
outputs will be latched off to protect against a
short−to−ground condition on SWDDQ or VDDQ. The IC
will be reset once VCCA or VDDQEN is cycled.
VTTREF Output
The VTTREF output tracks VDDQREF/2 at "2% accuracy.
It has source current capability of up to 15 mA. VTTREF
should be bypassed to analog ground of the device by
1.0 mF ceramic capacitor for stable operation. The VTTREF
is turned on as long as VDDQREF is pulled high. In S0 mode,
VTTREF soft−starts with VDDQ and tracks VDDQREF/2. In
S3 mode, VTTREF is kept on with VDDQ. VTTREF is turned
off only in S4/S5 like VDDQ output.
Feedback Compensation of VDDQ Regulator
The compensation network is shown in Figures 2 and 39.
VTT Active Terminator in Normal Mode (S0)
The VTT active terminator is a two−quadrant linear
regulator with two internal N−channel power FETs. It is
capable of sinking and sourcing at least 1.5 A continuous
current and up to 2.4 A transient peak current. It is activated
in normal mode in state S0 when the VTTEN pin is HIGH
and VDDQ is in regulation. Its input power path is from
VDDQ with the internal FETs gate drive power derived
from VCCA. The VTT internal reference voltage is derived
from the DDQREF pin. The VTT output is set to VDDQ/2
when VTT output is connecting to the FBVTT pin directly.
This regulator is stable with only a minimum 20 mF output
capacitor. The VTT regulator will have an internal
soft−start when it is transited from disable to enable.
During the VTT soft−start, a current limit is used as a current
source to charge up the VTT output capacitor. The current
limit is initially 1.0 A during VTT soft−start. It is then
increased to 2.5 A after 128 internal clock cycles which is
typically 0.32 ms.
Output Voltages Sensing
The VDDQ output voltage is sensed across the FBDDQ
and AGND pins. FBDDQ should be connected through a
feedback resistor divider to the VDDQ point of regulation
which is usually the local VDDQ bypass capacitor for load.
The AGND should be connected directly through a sense
trace to the remote ground sense point which is usually the
ground of local VDDQ bypass capacitor for load.
The VTT output voltage is sensed between the FBVTT
and VTTGND pins. The FBVTT should be connected to
the VTT regulation point, which is usually the VTT local
bypass capacitor, via a direct sense trace. The VTTGND
should be connected via a direct sense trace to the ground
of the VTT local bypass capacitor for load.
http://onsemi.com
16
NCP5218
Supply Voltage Undervoltage Monitor
MOSFET to discharge the excessive output voltage. When
the VDDQ output voltage goes back down to the nominal
regulation voltage, normal switching cycles are resumed.
When the VDDQ output exceeds 130% (typ) of the nominal
regulation voltage for 4 consecutive internal clock cycles,
the controller sets overvoltage fault, the device is latched
off by turning off both the high−side and low−side
MOSFETs. The overvoltage fault latch can be reset and the
controller can be restarted by toggling VDDQEN, VCCA,
or VIN.
The IC continuously monitors VCCA and VIN through
VCCA pin and OCDDQ pin respectively. VCCAGD is set
HIGH if VCCA is higher than its preset threshold (derived
from VREF with hysteresis). The IC will enter S5 state if
VCCA fails while in S0 and both VDDQEN and VTTEN
remain HIGH.
Thermal Shutdown
When the chip junction temperature exceeds 150_C, the
entire IC is shutdown. The IC resumes normal operation
only after the junction temperature dropping below 125_C.
Undervoltage Protection
In S3 power−saving mode with reduced switching at
lighter loads, when the VDDQ falls below 94% of the
nominal regulation voltage, the reduced switching
frequency is raised up back to the maximum switching
frequency. When VDDQ voltage is back to nominal
regulation voltage, the normal S3 power−saving operation
is resumed. In both S0 and S3 modes, when the VDDQ falls
below 65% (typ) of the nominal regulation voltage for 4
consecutive internal clock cycles, the undervoltage fault is
set, the device is latched off by turning off both the
high−side and low−side MOSFETs. The output is
discharged by the load current. The load current and output
capacitance determine the discharge rate. Cycling
VDDQEN, VCCA, or VIN can reset the undervoltage fault
latch and restart the controller.
Power Good
The PGOOD is an open−drain output of a window
comparator which continuously monitors the VDDQ
output voltage. The PGOOD is pulled low when the VDDQ
rises 12% above or drops 12% below the nominal
regulation point. The PGOOD becomes high impedance
when the VDDQ is within ±12% of the preset nominal
regulation voltage. A 100 kW resistor is recommended to
connect between PGOOD and VCCA as pull−up resistor
for logic level output.
Overvoltage Protection
When the VDDQ output is above 106% but below 130%
of the nominal regulation output voltage, the controller
turns off the high−side MOSFET and turns on the low−side
http://onsemi.com
17
NCP5218
VCCA
VIN
(VOCDDQ)
VDDQEN
VTTEN is
Don’t Care
in S5
VTTEN
VDDQ
VDDQ
Soft−start
VTT in H−Z
VTT
VTT Soft−start
VTT Soft−start
VTTREF
PGOOD
thold X 200 ms
Operating
Mode
S5
VCCA goes
above 4.0 V to
enable the IC.
VDDQEN goes HIGH,
VDDQ and VTTREF
are enabled but not
activated until VIN
goes above threshold
of 3.0 V. VTTEN goes
HIGH, VTT is enabled
but not activated until
VDDQ is good.
S0
PGOOD
goes HIGH.
S3
S0
VTTEN goes LOW
to activate S3 mode
and to turn off VTT.
INREGDDQ goes
HIGH, VTT goes into
normal mode.
VTTEN goes
HIGH, VTT goes
into normal mode.
VIN goes above the
threshold, the VDDQ
and VTTREF go into
normal mode.
Figure 38. Powerup and Powerdown Timing Diagram
http://onsemi.com
18
S5
Both VDDQEN and
VTTEN go LOW to
trigger S5 mode;
VDDQ, VTT, VTTREF
are disabled, then
INREGDDQ and
PGOOD goes LOW.
NCP5218
APPLICATION INFORMATION
Input Capacitor Selection for VDDQ Buck Regulator
Vripple + IL(ripple)
The input capacitor is important for proper regulation
operation of the buck regulator. It minimizes the input
voltage ripple and current ripple from the power source by
providing a local loop for switching current. The input
capacitor should be placed close to the drain of the
high−side MOSFET and source of the low−side MOSFET
with short, wide traces for connection. The input capacitor
must have large enough rms ripple current rating to
withstand the large current pulses present at the input of the
bulk regulator due to the switching current. The required
input capacitor rms ripple current rating can be estimated
by the following with minimum VIN:
ICIN(RMS) w IOUT
2
V
ǸVVOUT
* ǒ OUTǓ
V
IN
IN
(eq. 3)
where IL(ripple) is the inductor ripple current, ton is on−time,
and COUT is the output capacitance.
The inductor ripple current can be calculated by the
equation:
IL(ripple) +
(VIN−VOUT)
L fSW
VOUT
VIN
(eq. 4)
where L is the inductance and fSW is the switching
frequency. The output ripple voltage can be reduced by
either using the inductor with larger inductance or the
output capacitor with smaller ESR. Thus, the ESR needed
to meet the ripple voltage requirement can be obtained by:
(eq. 1)
ESR v
Besides, the voltage rating of the input capacitor should
be at least 1.25 times of the maximum input voltage.
Capacitance of around 20 mF to 50 mF should be sufficient
for most DDR applications. Ceramic capacitors are the
most suitable choice of input capacitor for notebook
applications due to their low ESR, high ripple current, and
high voltage rating. POSCAP or OS−CON capacitors can
also be used since they have good ESR and ripple current
rating, but they are larger in size and more expensive.
Aluminum electrolytic capacitors are also a choice for their
high voltage rating and low cost, but several aluminum
capacitors in parallel should be used for the required ripple
current. If ceramic capacitors are used, X5R and X7R types
are preferred rather than the Y5V type since the X5R and
X7R types are ceramic capacitors and have smaller
tolerance and temperature coefficient.
Vripple L fSW VIN
(VIN−VOUT) VOUT
(eq. 5)
The inductor ripple current is typically 30% of the
maximum load current and the ripple voltage is typically
2% of the output voltage. Thus, the above inequality can be
simplified to:
ESR v
0.02 VOUT
0.3 ILOAD(max)
(eq. 6)
For the load transient, the output capacitor contributes to
both the load−rise and the load−release responses. The
voltage undershoot during step−up load can be calculated
by the equation:
Vundershoot + DILOAD
ESR )
ǒ
Ǔ
1− VVOUT
IN
fSW
DILOAD
COUT
(eq. 7)
Output Capacitor Selection for VDDQ Buck Regulator
where DILOAD is the change in output current. If the second
term is ignored, then it becomes the following inequality:
The output filter capacitor plays an important role in
steady state output ripple voltage, load transient
requirement, and loop compensation stability. The ESR
and the capacitance of the output capacitor are the most
important parameters needed to be considered. In general,
the output capacitor must have small enough ESR for
output ripple voltage and load transient requirement.
Besides, the capacitance of the output capacitor should be
large enough to meet the overshoot and undershoot during
load transient. Since steady state output ripple voltage,
transient load undershoot and overshoot are the largest at
maximum VIN, the ESR and capacitance of output
capacitor should be estimated at the maximum VIN
condition.
For steady output ripple voltage, both ESR and
capacitance of the output capacitor are the contributing
factors, however, the capacitor ESR is the dominant factor.
The output ripple voltage is calculated as follows:
Vripple + IL(ripple)
ESR, for small ton and large COUT
V
ESR v undershoot
DILOAD
(eq. 8)
The maximum ESR requires to meet voltage undershoot
requirement at step−up load transient can be estimated
from the above inequality.
Then, the required output capacitor capacitance can be
obtained by the following:
DILOAD
COUT w
Vundershoot−DILOAD
ESR
ǒ
Ǔ
1− VVOUT
IN
fSW
(eq. 9)
The output voltage overshoot during load−release is
because the excessive stored energy in the inductor is
absorbed by the output capacitor. The overshoot voltage
can be calculated by the following equation:
Vovershoot +
IL(ripple) ton
(eq. 2)
ESR )
COUT
) COUTV2OUT
ǸLI2STEP(peak)COUT
−VOUT
(eq. 10)
http://onsemi.com
19
NCP5218
Then the required output capacitor capacitance can be
estimated by:
COUT w
L
I2STEP(peak)
(Vovershoot ) VOUT)2−V2OUT
ISTEP(peak) + DILOAD )
(VIN−VOUT)
2L fSW
IL(rating) + 1.2
(eq. 11)
IL(ripple)
2
(VIN−VOUT) VOUT
+ ILOAD(max) )
2 L fSW VIN
IL(peak) + ILOAD(max) )
VOUT
VIN
(eq. 12)
ICOUT(RMS) w IL(ripple) +
VOUT
(VIN−VOUT)
L fSW
(eq. 17)
Since the excessive energy stored in the inductor
contributed to the output voltage overshoot during load
release, the following inequality can be used to ensure that
the selected inductance value can meet the voltage
overshoot requirement at load release:
Lv
(eq. 13)
VOUT
VIN
(eq. 14)
Inductor Selection
The inductor should be chosen according to the inductor
ripple current, inductance, maximum current rating,
transient load release, and DCR.
In general, the inductor ripple current is 20% to 40% of
the maximum load current. A ripple current of 30% of the
maximum load current can be used as a typical value. The
required inductance can be estimated by:
0.3
(VIN−VOUT)
ILOAD(max)
VOUT
VIN fSW
COUT
((Vovershoot ) VOUT)2−V2OUT)
I2STEP(peak)
(eq. 18)
In addition, the inductor also needs to have low enough
DCR to obtain good conversion efficiency. In general,
inductors with about 2.0 mW to 3.0 mW per mH of
inductance can be used. Besides, larger inductance value
can be selected to achieve higher efficiency as long as it
still meets the targeted voltage overshoot at load release
and inductor DC current rating. Moreover, it should be
noted that using too small inductance value will cause very
large inductor ripple current in CCM in S0 mode and
extremely large peak inductor current in DCM in
power−saving mode during S3 mode. For both cases,
output capacitors with smaller ESR and larger capacitance
are required to keep the output ripple voltage small. It
should also be noted that the peak inductor current under
DCM light−load condition in power−saving mode in S3
mode will be larger than the peak inductor current under
heavy−load condition in S0 mode when very small
inductance value is used. Besides, using smaller
inductance will achieve lower efficiency
and
require larger minimum load to maintain nominal voltage
regulation in power−saving mode in S3 state. Therefore, it
is recommended that the inductance value should be at
least 0.56 mH or above to obtain optimum performance.
SP−Cap, POSCAP and OS−CON capacitors are suitable
for the output capacitor since their ESR is low enough to
meet the ripple voltage and load transient requirements.
Usually, two or more capacitors of the same type,
capacitance and ESR can be used in parallel to achieve the
required ESR and capacitance without change the ESR
zero position for maintaining the same loop stability. Other
than the performance point of view, the physical size and
cost are also the concerned factors for output capacitor
selection.
Lw
(eq. 16)
where IL(peak) is the peak inductor current at maximum load
current which is determined by:
where ISTEP(peak) is the load current step plus half of the
ripple current at the load release and DILOAD is the change
in the output load current.
Besides, the ESR and the capacitance of the output filter
capacitor also contribute to double pole and ESR zero
frequencies of the output filter, and the poles and zeros
frequencies of the compensation network for close loop
stability. The compensation network will be discussed in
more detail in the Loop Compensation section.
Other parameters about output filter capacitor that
needed to be considered are the voltage rating and ripple
current rating. The voltage rating should be at least
1.25 times the output voltage and the rms ripple current
rating should be greater than the inductor ripple current.
Thus, the voltage rating and ripple current rating can be
obtained by:
Vrating w 1.25
IL(peak)
MOSFET Selection
External N−channel MOSFETs are used as the switching
elements of the buck controller. Both high−side and
low−side MOSFETs must be logic−level MOSFETs which
can be fully turned on at 5.0 V gate−drive voltage.
On−resistance (RDS(on)), maximum drain−to−source
voltage (VDSS), maximum drain current rating, and gate
charges (QG, QGD, QGS) are the key parameters to be
considered when choosing the MOSFETs.
For on−resistance, it should be the lower; the better is the
performance in terms of efficiency and power dissipation.
Check the MOSFET’s rated RDS(on) at VGS = 4.5 Vwhen
selecting the MOSFETs. The low−side MOSFET should
have lower RDS(on) than the high−side MOSFET since the
(eq. 15)
where ILOAD(max) is the maximum load current.
The DC current rating of the inductor should be about 1.2
times of the peak inductor current at maximum output load
current. Therefore, the maximum DC current rating of the
inductor can be obtained by:
http://onsemi.com
20
NCP5218
VDDQ. i.e., the voltage drop across the RDS(on) of high−side
FET developed by the drain current is larger than the
voltage drop across RL1, the OCP is triggered and the
device will be latched off.
The overcurrent protection will trip when a peak inductor
current hit the ILIMIT determined by the equation:
turn−on time of the low−side MOSFET is much longer than
the high−side MOSFET in high VIN and low VOUT buck
converter. Generally, high−side MOSFET with RDS(on)
about 7.0 mW and low−side MOSFET with RDS(on) about
5.0 mW can achieve good efficiency.
The maximum drain current rating of the high−side
MOSFET and low−side MOSFET must be higher than the
peak inductor current at maximum load current. The
low−side MOSFET should have larger maximum drain
current rating than the high−side MOSFET since the
low−side MOSFET have longer turn−on time.
The maximum drain−to−source voltage rating of the
MOSFETs used in buck converter should be at least 1.2 times
of the maximum input voltage. Generally, VDSS of 30 V
should be sufficient for both high−side MOSFET and
low−side MOSFET of the buck converter for notebook
application.
As a general rule of thumb, the gate charges are the
smaller; the better is the MOSFET while RDS(on) is still low
enough. MOSFETs are susceptible to false turn−on under
high dV/dt and high VDS conditions. Under high dV/dt and
high VDS condition, current will flow through the CGD of
the capacitor divider formed by CGD and CGS, cause the
CGS to charge up and the VGS to rise. If the VGS rises above
the threshold voltage, the MOSFET will turn on.
Therefore, it should be checked that the low−side MOSFET
have low QGD to QGS ratio. This indicates that the low−side
MOSFET have better immunity to short moment false
turn−on due to high dV/dt during the turn−on of the
high−side MOSFET. Such short moment false turn−on will
cause minor shoot−through current which will degrade
efficiency, especially at high input voltage condition.
ILIMIT + RL1 IOC
RDS(on)
(eq. 19)
It should be noted that the OCDDQ pin must be pulled
high to VIN through a resistor RL1 and this pin cannot be
left floating for normal operation. The voltage drop across
RL1 must be less than 1.0 V to allow enough headroom for
the voltage detection at the OCDDQ pin under low VIN
condition. In addition, since the MOSFET RDS(on) varies
with temperature as current flows through the MOSFET
increases, the OCP trip point also varies with the MOSFET
RDS(on) temperature variation.
Since the IOC and RDS(on) have device variations and
MOSFET RDS(on) increase with temperature, to avoid false
triggering the overcurrent protection in normal operating
output load range, calculate the RL1 value from the
previous equation with the following conditions such that
minimum value of inductor current limit is set:
1. The minimum IOC value from the specification
table.
2. The maximum RDS(on) of the MOSFET used at
the highest junction temperature.
3. Determine ILIMIT for ILIMIT > ILOAD(max) +
IL(ripple)/2, where ILOAD(max) = IVDDQ(max) +
IVTT(max) if VTT is powered by VDDQ.
Besides, a decoupling capacitor CDCPL should be added
closed to the lead of the current limit setting resistor RL1
which connected to the drain of the high−side MOSFET.
Overcurrent Protection of VDDQ Buck Regulator
The OCP circuit is configured to set the current limit for
the current flowing through the high−side FET and
inductor during S0 and S3. The overcurrent tripping level
is programmed by an external resistor RL1 connected
between the OCDDQ pin and drain of the high−side FET.
An internal 31 mA current sink (IOC) at pin OCDDQ
establishes a voltage drop across the resistor RL1 at a
magnitude of RL1xIOC and develops a voltage at the
non−inverting input of the current limit comparator.
Another voltage drop is established across the high−side
MOSFET RDS(on) at a magnitude of ILxRDS(on) and a
voltage is developed at SWDDQ when the high−side
MOSFET is turned on and the inductor current flows
through the RDS(on) of the MOSFET. The voltage at the
non−inverting input of the current limit comparator is then
compared to the voltage at SWDDQ pin when the
high−side gate drive is high after a fixed period of blanking
time (150 ns) to avoid false current limit triggering. When
the voltage at SWDDQ is lower than the voltage at the
non−inverting input of the current limit comparator for four
consecutive internal clock cycles, an overcurrent condition
occurs, during which, all outputs will be latched off to
protect against a short−to−ground condition on SWDDQ or
Loop Compensation
Once the output LC filter components have been
determined, the compensation network components can be
selected. Since NCP5218 is a voltage mode PWM
converter with output LC filter, Type III compensation
network is required to obtain the desired close loop
bandwidth and phase boost with unconditional stability.
The NCP5218 PWM modulator, output LC filter and
Type III compensation network are shown in Figure 39.
The output LC filter has a double pole and a single zero.
The double pole is due to the inductance of the inductor and
capacitance of the output capacitor, while the single zero
is due to the ESR and capacitance of the output capacitor.
The Type III compensation has two RC pole−zero pairs.
The two zeros are used to compensate the LC double pole
and provide 180° phase boost. The two poles are used to
compensate the ESR zero and provide controlled gain
roll−off. For an ideally compensated system, the Bode plot
should have the close−loop gain roll−off with a slope of
−20 dB/decade crossing the 0 dB with the required
bandwidth and the phase margin larger than 45° for all
frequencies below the 0 dB frequency. The closed loop
http://onsemi.com
21
NCP5218
having −20 dB/decade slope at 0 dB gain crossing, and
sufficient phase margin for stability. The bandwidth of
close loop gain should be less than 50% of the switching
frequency and the compensation gain should be bounded
by the error amplifier open loop gain.
gain is obtained by adding the modulator and filter gain (in
dB) to the compensation gain (in dB).The bandwidth is the
frequency at which the gain is 0 dB and the phase margin
is the difference between the close loop phase and 180°.
The goal of compensation is to achieve a stable close loop
system with the highest possible bandwidth, the gain
VIN
NCP5218
CIN
VBOOST
Q1
TGDDQ
VDDQ
PWM
LOGIC
SWDDQ
VCCP
L
VDDQ
Q2
ESR
BGDDQ
COUT
PGND
OUTPUT
FILTER
PGND
OSC
PWM
COMP
COMP
ERROR
AMP
VIN
ADAPTIVE VRAMP
RAMP
C2
VREF
C1
R3
A
C3
R4
COMPENSATION
NETWORK
R1
FBDDQ
MODULATOR
R2
Figure 39. Voltage Mode Buck Converter with Modulator, LC filter and Type III Compensation
Modulator DC Gain can be calculated by:
VIN
GMOD(DC) + 20 log
VRAMP
Type III compensation poles and zeros break frequencies
are defined by the below equations:
(eq. 20)
fZ1 +
2p
LC filter double pole and ESR zero break frequencies are
defined by:
fPLC +
2p
fZESR +
2p
ǸL
1
(eq. 21)
COUT
1
ESR
fP1 +
COUT
(eq. 23)
100
fZ1
fZ2
80
fP1
fP2
GAIN (dB)
60
Open Loop Error
Amp Gain
40
Compensation
Gain
20
0
20 log
−20
−40
−60
20 log
10
R3
R1
VIN
VRAMP
100
Closed Loop Gain
Modulator & Filter Gain
fZESR
fPLC
1k
R3
10 k 100 k 1 M
10 M
FREQUENCY (Hz)
Figure 40. Asymptotic Bode Plot of the Converter Gain
http://onsemi.com
22
1
R4
(eq. 24)
C2
ǒCC11)CC22Ǔ
1
(R1 ) R4)
fP2 +
2p
Compensation network DC Gain can be calculated by the
equation:
R
GCOMP(DC) + 20 log 3
R1
1
2p
fZ2 +
2p
(eq. 22)
1
R3
C3
C3
(eq. 25)
(eq. 26)
(eq. 27)
NCP5218
By using the above equations and guidelines, the
compensation components values can be determined by the
equations below:
Close loop system bandwidth can be calculated by:
BW +
R3
R1
VIN
VRAMP
2p
ǸL
1
COUT
(eq. 28)
Since the ramp amplitude of the PWM modulator has a
voltage feedforward function, the ramp amplitude is a
function of VIN which can be determined by:
VRAMP + 1.25 V ) 0.045
R3 +
2p
BW
C2 +
(VIN−5.0 V) (eq. 29)
Below are some guidelines for setting the compensation
components:
1. Set a value for R1 between 2.0 kW and 5.0 kW.
2. Set a target for the close loop bandwidth which
should be less than 50% and is typically 1/8 to
1/4 of the switching frequency.
3. Pick compensation DC gain (R3/R1) for desired
close loop bandwidth.
4. Place 1st zero at half filter double pole.
5. Place 1st pole at ESR zero.
6. Place 2nd zero at filter double pole.
7. Place 2nd pole at half the switching frequency.
ESR
COUT (eq. 30)
ǸL C
OUT
R3
(eq. 31)
p
C2
2 Ǔ*1
ǒESRR3 CCOUT
R1
ǸL C
OUT * 1
fSW
1
R4
C3 +
p
fSW
(eq. 32)
(eq. 33)
(eq. 34)
The modulator and filter gain, compensation gain, and
close loop gain asymptotic Bode plot can be drawn by the
calculated results to check the compensation gain and close
loop gain obtained. An example of asymptotic Bode plot is
shown in Figure 40.
The phase of the output filter can be calculated by:
Phase(Filter) + − tan −1(2pf
2
C1 +
R4 +
ǸL
VRAMP R1
VIN
COUT)− tan −1
) DCR
ǒ2pf(2pfESR
2
)
L C
COUT
OUT−1
Ǔ
(eq. 35)
where the DCR of the inductor can be neglected if the DCR is small.
The phase of the Type III compensation network can be calculated by:
Phase(TypeIII) + −90° ) tan −1(2pf
) tan −1(2pf
ǒ
C2)− tan −1 2pf
R3
(R1 ) R4)
C3)− tan −1(2pf
C1 C2
C1 ) C2
R4 C3)
R3
Ǔ
(eq. 36)
The close loop phase can be calculated by summing the
filter phase and compensation phase:
0.8 R1
R2 +
VOUT−0.8
Phase(CloseLoop) + Phase(Filter) ) Phase(TypeIII)
It is recommended to adjust the value of R2 to fine−tune
the output voltage when it is necessary. The value of R1
should not be changed since the compensation DC gain and
the 2nd zero break frequency of the compensation gain are
contributed by R1. If the value of R1 is changed, the
compensation, the close loop bandwidth and phase margin,
and the system stability will be affected. Besides, it is
recommended to use resistors with at least 1% tolerance for
R1 and R2.
(eq. 37)
Then the close loop phase margin can be estimated by:
Phase(Margin) + Phase(CloseLoop) * (*180°)
(eq. 38)
It should be checked that closed loop gain has a 0 dB gain
crossing with −20 dB/decade slope and a phase margin of
45° or greater. The compensation components values may
require some adjustment to meet these requirements.
Besides, the compensation gain should be checked with the
error amplifier open loop gain to make sure that it is
bounded by the error amplifier open loop gain.
The poles and zeros locations and hence the
compensation network components values may need to be
further fine tuned after actual system testing and analysis.
(eq. 39)
Soft−Start of Buck Regulator
A VDDQ soft−start feature is incorporated in the device
to prevent surge current from power supply and output
voltage overshoot during power up. When VDDQEN,
VCCA, and VOCDDQ rise above their respective upper
threshold voltages, the external soft−start capacitor CSS
will be charged up by a constant current source, Iss. When
the soft−start voltage (Vcss) rises above the SS_EN voltage
(X50 mV), the BGDDQ and TGDDQ will start switching
and VDDQ output will ramp up with VFBDDQ following
the soft−start voltage. When the soft−start voltage reaches
the SS_OK voltage (XVREF + 50 mV), the soft−start of
Feedback Resistor Divider
The output voltage of the buck regulator can be adjusted
by the feedback resistor divider formed by R1 and R2. Once
the value of R1 is selected when determining the
compensation components, the value of R2 can be obtained
by:
http://onsemi.com
23
NCP5218
VDDQ is finished. The Css will continue to charge up until
it reaches about 2.5 V to 3.0 V.
The soft−start time tss can be programmed by the
soft−start capacitor according to the following equation:
tss [
0.8
Css
Iss
to VTTGND with at least a 10 mF capacitor if external
voltage source is used.
Design Example
A design example of a VDDQ bulk converter with the
following design parameters is shown below:
(eq. 40)
DDR2 VDDQ bulk converter design parameters:
1. Input voltage range: 7.0 V to 20 V.
2. Nominal VOUT: 1.8 V.
3. Static tolerance: 2% ("36 mV).
4. Transient tolerance: "100 mV.
5. Maximum output current: 10 A
(IVDDQ(max) = 8.0 A, IVTT(max) = 2.0 A).
6. Load transient step: 1.0 A to 8.0 A.
7. Switching frequency: 400 kHz.
8. Bandwidth: 100 kHz.
9. Soft−start time: 400 ms.
a. Calculate input capacitor rms ripple current rating and
voltage rating:
Ceramic capacitors with low tolerance and low
temperature coefficient, such as B, X5R, X7R ceramic
capacitors are recommended to be used as the CSS. Ceramic
capacitors with Y5V temperature characteristic are not
recommended.
Soft−Start of VTT Active Terminator
The VTT source current limit is used as a constant
current source to charge up the VTT output capacitor
during VTT soft−start. Besides, the VTT source current
limit is reduced to about 1.0 A for 128 internal clock cycles
to minimize the inrush current during VTT soft−start.
Therefore, the VTT soft−start time tSSVTT can be estimated
by the equation:
C
VTT
tSSVTT [ OUTVTT
ILIMVTSS
ICIN(RMS) w 10 A
(eq. 41)
V * ǒ1.836 VǓ2 + 4.2 A
Ǹ1.836
8.0 V
8.0 V
(eq. 42)
where COUTVTT is the capacitance of VTT output capacitor
and ILIMVTSS is the VTT soft−start source current limit.
VCIN(rating) w 20
1.25 V + 25 V
(eq. 43)
Therefore, two 10 mF 25 V ceramic capacitors with 1210
size in parallel are used.
b. Calculate inductance, rated current and DCR of
inductor:
First, suppose ripple current is 0.3 times the maximum
output current, such that:
Boost Supply Diode and Capacitor
An external diode and capacitor are used to generate the
boost voltage for the supply of the high−side gate driver of
the bulk regulator. Schottky diode with low forward
voltage should be used to ensure higher floating gate drive
voltage can be applied across the gate and the source of the
high−side MOSFET. A Schottky diode with 30 V reverse
voltage and 0.5 A DC current ratings can be used as the
boost supply diode for most applications. A 0.1 mF to
0.22 mF ceramic capacitor should be sufficient as the boost
capacitor.
Lw
(20 V−1.836 V) 1.836 V
+ 1.39 mH (eq. 44)
0.3 10 A 20 V 400 kHz
Second, the overshoot requirement at load release is then
considered and supposes two 220 mF capacitors in parallel
are used as an initially guess, such that:
440 mF
Lv
VTTI Input Power Supply for VTT and VTTR
((100 mV )1.836 V)2−(1.836 V)2)
+2.56 mH
ǒ7 A ) 0.3 7 AǓ2
2
Both VTT and VTTR are supplied by VTTI for sourcing
current. VTTI is normally connected to the VDDQ output
for optimum performance. If VTTI is connected to VDDQ,
no bypass capacitor is required to add to VTTI since the
bulk capacitor at VDDQ output is sufficiently large.
Besides, the maximum load current of VDDQ is the sum of
IVDDQ(max) and IVTT(max) when making electrical design
and components selection of the VDDQ buck regulator.
VTTI can also be connected to an external voltage source.
However, extra power dissipation will be generated from
the internal VTT high−side MOSFET and more
heatsinking is required if the external voltage is higher than
VDDQ. Whereas, the headroom will be limit by the RDS(on)
of the VTT linear regulator high−side MOSFET, and the
maximum VTT output current with VTT within regulation
window will also be reduced if the external voltage is lower
than VDDQ. Besides, the VTTI pin input must be bypassed
(eq. 45)
Thus, inductors with standard inductance values of
1.5 mH, 1.8 mH and 2.2 mH can be used. As a trade−off
between smaller overshoot and better efficiency, the
average value of 1.8 mH inductor is selected.
Then, the maximum rated DC current is calculated by:
IL(rated) + 1.2
V−1.836 V) 1.836 V
ǒ10 A ) 2(201.8
Ǔ
mH 400 kHz 20
+ 13.39 A
(eq. 46)
Therefore, inductor with maximum rated DC current of
14 A or larger can be used.
Finally, the DCR of inductor is 2.0 mW per mH of
inductance as a rule of thumb, then:
DCR + 2 mW
1 mH
http://onsemi.com
24
1.8 mH + 3.6 mW
(eq. 47)
NCP5218
Thus, inductor with 1.8 mH inductance, 14 A maximum
rated DC current and 3.5 mW DCR is chosen.
c. Calculate ESR and capacitance of output filter
capacitor:
First, the ESR required to achieve the desired output
ripple voltage is considered. Suppose the output ripple
voltage is 2% of the nominal output voltage.
1.8 V) 1.8 mH
(20 V−1.8 V)
+ 15.8 mW
ESR v
(0.02
400 kHz
1.8 V
Second, the ESR required to meet the transient load
undershoot requirement is considered, such that:
ESR v 100 mV + 14.3 mW
7A
(eq. 49)
Therefore, the suitable ESR is 12 mW or smaller, and the
value of 7.5 mW is selected for more design margin and
better performance. Then, two same SP−Caps or POSCAPs
each with 15 mW ESR in parallel having a resultant ESR
of 7.5 mW should be good enough to meet the
requirements.
Then, check that whether the previously supposed
capacitance meets the undershoot and overshoot
requirements.
20 V
(eq. 48)
To ensure that undershoot requirement of less than 100 mV is achieved, the capacitance must be:
7A
COUT w
100 mV−7 A
ǒ
Ǔ
mVǓ
ǒ1− 1.8 V−36
20 V
7.5 mW
400 kHz
(eq. 50)
+ 335.9 mF
To make sure that overshoot requirement of less than 100 mV is fulfilled, capacitance must be:
COUT w
1.8 mH
V) 1.836 V 2
ǒ7 A ) 2(201.8V−1.836
Ǔ
mH 400 kHz 20 V
(100 mV ) 1.836 V)2 − (1.836 V)2
Therefore, output capacitor with capacitance of 440 mF
should meet both undershoot and overshoot requirements.
Sometimes, it may take several times of iterations between
the process of selecting inductance of the inductor and ESR
and capacitance of the output capacitor.
Then, the voltage rating of the output capacitor is
estimated by:
Vrated w 1.25
1.836 V + 2.3 V
(eq. 51)
+ 317.6 mF
d. Calculate the resistance value of OCP current limit
setting resistor:
First, the OCP current limit is estimated at maximum
load condition, such that:
ILIMIT u 8 A ) 2 A )
2
+ 11.16 A
(eq. 52)
(20 V−1.836 V) 1.836 V
1.8 mH 400 kHz 20 V
(eq. 54)
Thus, ILIMIT is set to 11.5 A. Suppose from the high−side
MOSFET data sheet, the maximum RDS(on) is 10 mW.
Then, the value of RL1 is calculated by:
Thus, output capacitor with 2.5 V or larger rated voltage
is used.
Finally, the rated rms ripple current of the output
capacitor is considered:
RL1 + 11.5 A 10 mW + 4.4 kW
26 mA
(20 V−1.836 V) 1.836 V
ICOUT(rms) w
+ 2.3 A
1.8 mH 400 kHz 20 V
(eq. 55)
Therefore, the resistor with standard value of 4.7 kW is
selected for RL1.
e. Calculate the RC values of the compensation network:
First, 4.3 kW is chosen as the value of R1 which is in the
range between 2.0 kW and 5.0 kW.
Since the worst case of stability is at the maximum VIN,
the close loop compensation should be considered at
maximum VIN. Then the ramp amplitude can be calculated
as below:
(eq. 53)
Thus, capacitor with rated rms ripple current of 3.0 A or
larger should be selected. Two capacitors each with 1.5 A
rated ripple current can be connected in parallel to provide
a total of 3.0 A rated rms ripple current.
Therefore, two same capacitors in parallel each with
capacitance of 220 mF, ESR of 15 mW, rated voltage of
2.5 V, and rated rms ripple current of 1.5 A are used.
VRAMP + 1.25 V ) 0.045
(20 V−5 V) + 1.925 V
(eq. 56)
Since the L = 1.8 mH, COUT = 440 mF, and the target close loop bandwidth is 100 kHz, the value of R3 can be
calculated as:
R3 +
2p
100 kHz
1.925 V
4.3 kW
20 V
Ǹ1.8 mH
http://onsemi.com
25
440 mF
+ 7.3 kW
(eq. 57)
NCP5218
Then, if the second zero break frequency is placed at the LC
filter’s double pole and the second pole is placed at half the
switching frequency, the value of R4 can be calculated by:
Thus, standard value of 7.5 kW is selected for R3.
If the first zero break frequency is placed at half the LC
filter’s double pole, the value of C2 can be calculated by:
C2 +
2
Ǹ1.8 mH
7.5 kW
440 mF
R4 +
(eq. 58)
+ 7.5 nF
Thus, standard value of 8.2 nF is chosen for C2.
If the 1st pole break frequency is placed at the LC filter’s
ESR zero, the value of C1 can be calculated by:
p
400 kHz
4.3 kW
Ǹ1.8 mH
440 mF −1
+ 125 W
(eq. 60)
Thus, standard value of 130 W is selected for R4.
Then, C3 can be calculated by:
C3 +
p
8.2 nF
C1 +
+ 464.9 pF
(eq. 59)
7.5 kW 8.2 nF * 1
7.5 mW 440 mF
1
+ 6.12 nF (eq. 61)
130 W 400 kHz
Therefore, standard value of 5.6 nF is selected for C3.
Thus, standard value of 470 pF can be chosen for C1.
However, 180 pF is selected for more phase boost at the
0 dB gain crossing.
Then, the close loop phase margin can be estimated by the following:
Phase(Filter) + − tan −1(2p
− tan −1
ǒ(2p
100 kHz
7.5 mW
440 mF)
2p 100 kHz 7.5 mW
100 kHz)2 1.8 mH 440 mF−1
Ǔ
+ −150.47°
Phase(TypeIII) + −90 ) tan −1(2p
ǒ
100 kHz
− tan −1 2p
100 kHz
) tan −1(2p
100 kHz
− tan −1(2p
100 kHz
7.5 kW 8.2 nF)
180 pF 8.2 nF
7.5 kW
180 pF ) 8.2 nF
(4.3 kW ) 130 W)
130 W
Ǔ
(eq. 62)
5.6 nF)
5.6 nF)
+ 20.57°
Phase(closeloop) + −150.47° ) 20.57° + −129.90°
Phase(margin) + Phase(closeloop)−(−180°) + −129.90°−(−180°) + 50.10°
Therefore, the phase margin is large enough for stability.
f. Calculate the resistance value of feedback resistor
divider:
Since a 4.3 kW resistor is chosen as the high−side resistor
R1, the resistance value of low−side resistor R2 can be
calculated by:
Therefore, a 3.44 kW resistor is selected for the low−side
feedback resistor R2.
g. Calculate soft−start capacitor value for the desired
400 ms VDDQ soft−start time:
R2 + 0.8 4.3 kW + 3.44 kW
1.8 V−0.8 V
Therefore, 2.0 nF X5R ceramic capacitor is selected for
the soft−start capacitor.
CSS +
(eq. 63)
http://onsemi.com
26
4.0 mA 400 ms
+ 2.0 nF
0.8 V
(eq. 64)
NCP5218
PCB Layout Guidelines
Cautious PCB layout design is very critical to ensure
high performance and stable operation of the DDR power
controller. The following items must be considered when
preparing PCB layout:
1. All high−current traces must be kept as short and
wide as possible to reduce power loss.
High−current traces are the trace from the input
voltage terminal to the drain of the high−side
MOSFET, the trace from the source of the
high−side MOSFET to the inductor, the trace
from inductor to the VDDQ output terminal, the
trace from the input ground terminal to the VDDQ
output ground terminal, the trace from VDDQ
output to VTTI pin, the trace from VTT pin to
VTT output terminal, and the trace from VTT
output ground terminal to the VTTGND pin.
Power handling and heaksinking of high−current
traces can be improved by also routing the same
high−current traces in the other layers and joined
together with multiple vias.
2. Power components which include the input
capacitor, high−side MOSFET, low−side
MOSFET and VDDQ output capacitor of the buck
converter section must be positioned close
together to minimize the current loop. The input
capacitor must be placed close to the drain of the
high−side MOSFET and the source of the
low−side MOSFET.
3. To ensure the proper function of the device,
separated ground connections should be used for
different parts of the application circuit according
to their functions. The input capacitor ground, the
low−side MOSFET source, the VDDQ output
capacitor ground, the VCCP decoupling capacitor
ground should be connected to the PGND. The
trace path connecting the source of the low−side
MOSFET and PGND pin should be minimized.
The VTT output capacitor ground should be
connected to the VTTGND first with a short
trace, it is then connected to the ground plane of
PGND. The VCCA decoupling capacitor ground,
the ground of the VDDQ feedback resistor, the
soft−start capacitor ground, the VTTREF output
capacitor ground should be connected to the
AGND. The AGND pin is then connected directly
through a sense trace to the remote ground sense
point of the PGND, which is usually the ground
of the local bypass capacitor for the load. Never
connect the AGND, PGND and VTTGND
together just under the thermal pad.
4. The thermal pad of the DFN22 package should be
connected to the ground planes in the internal
layer and bottom layer from the copper pad at top
layer underneath the package through six to eight
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
http://onsemi.com
27
vias with 0.6 mm hole−diameter to help heat
dissipation and ensure good thermal capability. It
is recommended to use PCB with 1 oz or 2 oz
copper foil. The thermal pad can be connected to
either PGND ground plane or AGND ground
plane but not both.
The input capacitor ground terminal, the VDDQ
output capacitor ground terminal and the source
of the low−side MOSFET must be connected to
the PGND ground plane through multiple vias.
Sensitive traces like trace from FBDDQ, trace
from COMP, trace from OCDDQ, trace from
FBVTT and trace from VTTREF should be
avoided from the high−voltage switching nodes
like SWDDQ, BOOST, TGDDQ and BGDDQ.
Separate sense trace should be used to connect
the VDDQ point of regulation, which is usually
the local bypass capacitor for load, to the
feedback resistor divider to ensure accurate
voltage sensing. The feedback resistor divider
should be place close to the FBDDQ pin.
Separate sense trace should be used to connect the
VTT point of regulation, which is usually the local
bypass capacitor for load, to the FBVTT pin.
Separate sense trace should be used to connect
the VDDQ point of regulation to the DDQREF pin
to ensure that the reference voltage to VTT is
accurately half of the VDDQ voltage.
The traces length between the gate driver outputs
and gates of the MOSFETs must be minimized to
avoid parasitic impedance.
To ensure normal function of the device, an RC
filter should be placed close to the VCCA pin and
a decoupling capacitor should be placed close to
the VCCP pin.
The copper trace area of the switching node which
includes the source of the high−side MOSFET,
drain of the low−side MOSFET and high voltage
side of the inductor should be minimized by using
short wide trace to reduce EMI.
A snubber circuit consists of a 3.3 W resistor and
1.0 nF capacitor may need to be connected across
the switching node and PGND to reduce the
high−frequency ringing occurring at the rising
edge of the switching waveform to obtain more
accurate inductor current limit sensing of the
VDDQ buck converter. However, adding this
snubber circuit will slightly reduce the conversion
efficiency.
VTTI should be connected to VDDQ output with
wide and short trace if VDDQ is used as the
sourcing supply for VTT. An input capacitor of at
least 10 mF should be added close to the VTTI
pin and bypassed to VTTGND if external voltage
supply is used as the VTT sourcing supply.
NCP5218
VCCA
U1
NCP5218
R1
R2
R3
R4
100 k 100 k 100 k 100 k
1
2
JP1
JP2
0.1 mF
VDDQEN OCDDQ
VTTEN
VCCP
FPWM
4 SS
BOOST
JP3
3
16
C1
1.8 nF ON Semiconductor
NCP5218
TP1
PGOOD
TP4
VREF
0.9 V/15 mA
1.25 V/15 mA TP10
AGND
TP2
VTT
0.9 V/±1.5 A
C2
1.25 V/±1.5 A
TP3 10 mF
VTTGND
5V
R5
R6
20
17
C17
10 mF
10 W
C3
10 mF
FBVTT
5 VTTGND
11
VCCA
4.7 mF
MBR0530T1
C6
2
1
D1
NTMS4700N
C7
Q1
0.1 mF
BGDDQ 21
22
PGND
12
COMP
C4
1 mF
10
13
DDQREF FBDDQ
(option)
9
AGND
VTTI
THPAD
23
5 V TP5
BIAS SUPPLY
5.6 kW
15 PGOOD TGDDQ 18
14 VTTREF
19
C18
SWDDQ
1 mF
6
VTT
8
C5 (option)
7
R7
0W
C8
10 mF
Q2
R8
0W
N−CHANNEL
C14
30 V, 4.7 mW
100 pF
R9
2.2 nF
10 k
L1
(option)
R14 C11
3.3 W 150 mF
C19
(option)
1 nF
C16
4.7 nF
R11
4.3 k
R10
130
* Install R12 = 3.44 k for VDDQ = 1.8 V
Install R12 = 2.02 k for VDDQ = 2.5 V
C20
10 mF
(option)
C9
*33 mF
(option for Vin < 8 V)
1.8 mH, 14 A, 3.4 mW
N−CHANNEL
30 V, 7.3 mW
NTMS4107N
C15
*
R12
3.44 k
JP4
VTTGND
R13
0W
Figure 41. Schematic Diagram of Evaluation Board
http://onsemi.com
28
TP6 VIN
(4.5 V TO 24 V)
C10
10 mF TP7
GND
VDDQ
C13
C12
1 mF
150 mF
TP8
VDDQ
1.8 V/10 A
2.5 V/12 A
TP9
VDDQGND
C11, C12
(150 mF, 4 V, 15 mW)
LOW ESR SP−CAP UD Series
Panasonic EEFUD0G151R
(150 mF, 4 V, 18 mW)
LOW−ESR POSCAP TPE Series
SANYO 4TPE150MI
NCP5218
PCB Layout of Evaluation Board
Figure 42. Silkscreen of Evaluation Board PCB
Figure 43. Top Layer of Evaluation Board
PCB Layout
Figure 45. Middle Layer2 of Evaluation
Board PCB Layout
Figure 44. Middle Layer1 of Evaluation
Board PCB Layout
Figure 46. Bottom Layer of Evaluation Board
PCB Layout
http://onsemi.com
29
NCP5218
Table 2. Bill of Materials of the Evaluation Board
Item
Qty
Designators
Part Description
Mfg. & P/N
1
1
C1
Capacitor, Ceramic, 1.8 nF/50 V 0603
Panasonic ECJ1VB1H182K
2
2
C2, C17
Capacitor, Ceramic, 10 μF/6.3 V 0805
Panasonic ECJ2FB0J106M
3
2
C3, C20
Capacitor, Ceramic, 10 μF/6.3 V 0805
Panasonic ECJ2FB0J106M
4
3
C4, C13, C18
Capacitor, Ceramic, 1 μF/10 V 0805
Panasonic ECJ1VB1A105M
5
2
C5, C7
Capacitor, Ceramic, 0.1 μF/25 V 0603
Panasonic ECJ1VB1E104K
6
1
C6
Capacitor, Ceramic, 4.7 μF/10 V 0603
Panasonic ECJ2FB1C475M
7
2
C8, C10
Capacitor, Ceramic, 10 μF/25 V 1210
Panasonic ECJ4YB1E106M
8
1
C9
Capacitor, Electrolytic, 33 μF/35 V Size D
Panasonic EEVFK1V330P
9
2
C11, C12
Capacitor, SP−CAP, 150 μF/4 V Size D /
Capacitor, POSCAP, 150 μF/4 V Size D
Panasonic EEFUD0G151R /
Sanyo 4TPE150MI
10
1
C14
Capacitor, Ceramic, 100 pF/50 V 0603
Panasonic ECJ1VC1H101K
11
1
C15
Capacitor, Ceramic, 2.2 nF/50 V 0603
Panasonic ECJ1VB1H222K
12
1
C16
Capacitor, Ceramic, 4.7 nF/50 V 0603
Panasonic ECJ1VB1H472K
13
1
C19
Capacitor, Ceramic, 1 nF/50 V 0603
Panasonic ECJ1VB1H102K
14
1
D1
Diode, 0.5 A 30 V schottky SOD−123
ON Semiconductor MBR0530T1
15
3
JP1, JP2, JP3
Header, 3−pin, 100 mil spacing
Any
16
1
JP4
Header, 2−pin, 100 mil spacing
Any
17
1
L1
Inductor, SMD, 1.8 μH/14 A /
Inductor, SMD, 1.5 μH/17 A
Panasonic ETQP2H1R8BFA /
TOKO FDA1055−1R5M=P3
18
1
Q1
MOSFET, N−Channel SO−8, 30 V/14.5 A
ON Semiconductor NTMS4700N
19
1
Q2
MOSFET, N−Channel SO−8, 30 V/19 A
ON Semiconductor NTMS4107N
20
4
R1, R2, R3, R4
Resistor, 100 kW 5% 0603
Panasonic ERJ3GEYJ104V
21
1
R5
Resistor, 10 W 5% 0603
Panasonic ERJ3GEYJ100V
22
1
R6
Resistor, 5.6 kW 1% 0603
Panasonic ERJ3EKF5602V
23
1
R7
Resistor, 0 W 5% 0603
Panasonic ERJ3GEYJ0R0V
24
2
R8, R13
Resistor, 0 W 5% 0603
Panasonic ERJ3GEYJ0R0V
25
1
R9
Resistor, 10 kW 1% 0603
Panasonic ERJ3EKF1002V
26
1
R10
Resistor, 130 W 1% 0603
Panasonic ERJ3EKF1300V
27
1
R11
Resistor, 4.3 kW 1% 0603
Panasonic ERJ3EKF4301V
28
1
R12
Resistor, 3.44 kW 1% 0603
Panasonic ERJ3EKF3441V
29
1
R14
Resistor, 3.3 W 5% 0603
Panasonic ERJ3GEYJ3R3V
30
8
TP1 − TP8
Header, single pin
Any
31
1
U1
2−in−1 Notebook DDR Power Controller
ON Semiconductor NCP5218
32
4
Shunt, 100 mil jumper
Any
33
1
Test Pin, 0.7 mm Diameter, 12 mm Height
Any
34
4
Bumpon, 4.44 x 0.20 transparent
3M
35
1
4−layered PCB 2500 mil x 2000 mil
Any
http://onsemi.com
30
Remark
C3 & C20 are
optional
C5 is optional
C9 is optional
C19 is optional
R14 is optional
Place at the
GND between
C11 and C8
NCP5218
PACKAGE DIMENSIONS
DFN22
MN SUFFIX
CASE 506AF−01
ISSUE A
A
D
B
PIN 1 LOCATION
NOTES:
1. DIMENSIONS AND TOLERANCING PER
ASME Y14.5M, 1994.
2. DIMENSIONS IN MILLIMETERS.
3. DIMENSION b APPLIES TO PLATED
TERMINALS AND IS MEASURED BETWEEN
0.25 AND 0.30 MM FROM TERMINAL
4. COPLANARITY APPLIES TO THE EXPOSED
PAD AS WELL AS THE TERMINALS.
E
0.15 C
DIM
A
A1
A3
b
D
D2
E
E2
e
K
L
TOP VIEW
0.15 C
0.10 C
A
0.08 C
SIDE VIEW
A1
(A3)
D2
22 X
L
1
e
C
SEATING
PLANE
MILLIMETERS
MIN
MAX
0.80
1.00
0.00
0.05
0.20 REF
0.18
0.30
6.00 BSC
3.98
4.28
5.00 BSC
2.98
3.28
0.50 BSC
0.20
−−−
0.50
0.60
SOLDERING FOOTPRINT*
11
4.300
0.169
22 X
0.980
0.039
E2
K
22
22 X
5.770
0.227
12
b
0.10 C A B
0.05 C NOTE 3
3.130
0.123
0.340
0.013
BOTTOM VIEW
0.500
20X 0.020
0.280
22X 0.011
SCALE 8:1
mm Ǔ
ǒinches
*For additional information on our Pb−Free strategy and soldering
details, please download the ON Semiconductor Soldering and
Mounting Techniques Reference Manual, SOLDERRM/D.
ON Semiconductor and
are registered trademarks of Semiconductor Components Industries, LLC (SCILLC). SCILLC reserves the right to make changes without further notice
to any products herein. SCILLC makes no warranty, representation or guarantee regarding the suitability of its products for any particular purpose, nor does SCILLC assume any
liability arising out of the application or use of any product or circuit, and specifically disclaims any and all liability, including without limitation special, consequential or incidental
damages. “Typical” parameters which may be provided in SCILLC data sheets and/or specifications can and do vary in different applications and actual performance may vary over
time. All operating parameters, including “Typicals” must be validated for each customer application by customer’s technical experts. SCILLC does not convey any license under
its patent rights nor the rights of others. SCILLC products are not designed, intended, or authorized for use as components in systems intended for surgical implant into the body,
or other applications intended to support or sustain life, or for any other application in which the failure of the SCILLC product could create a situation where personal injury or death
may occur. Should Buyer purchase or use SCILLC products for any such unintended or unauthorized application, Buyer shall indemnify and hold SCILLC and its officers, employees,
subsidiaries, affiliates, and distributors harmless against all claims, costs, damages, and expenses, and reasonable attorney fees arising out of, directly or indirectly, any claim of
personal injury or death associated with such unintended or unauthorized use, even if such claim alleges that SCILLC was negligent regarding the design or manufacture of the part.
SCILLC is an Equal Opportunity/Affirmative Action Employer. This literature is subject to all applicable copyright laws and is not for resale in any manner.
PUBLICATION ORDERING INFORMATION
LITERATURE FULFILLMENT:
Literature Distribution Center for ON Semiconductor
P.O. Box 5163, Denver, Colorado 80217 USA
Phone: 303−675−2175 or 800−344−3860 Toll Free USA/Canada
Fax: 303−675−2176 or 800−344−3867 Toll Free USA/Canada
Email: [email protected]
N. American Technical Support: 800−282−9855 Toll Free
USA/Canada
Europe, Middle East and Africa Technical Support:
Phone: 421 33 790 2910
Japan Customer Focus Center
Phone: 81−3−5773−3850
http://onsemi.com
31
ON Semiconductor Website: www.onsemi.com
Order Literature: http://www.onsemi.com/orderlit
For additional information, please contact your local
Sales Representative
NCP5218/D