FAIRCHILD AN-53

www.fairchildsemi.com
Application Note 53
Implementing an RC5051 DC-DC Converter on
Pentium® II Motherboards
Introduction
This document describes how to implement a switching voltage regulator using an RC5051 high speed controller, a
power inductor, a Schottky diode, appropriate capacitors,
and external power MOSFETs. This regulator forms a step
down DC-DC converter that can deliver up to 14.5A of
continuous load current at voltages ranging from 1.3V to
3.5V. A specific application circuit, design considerations,
component selection, PCB layout guidelines, and performance evaluations are covered in detail.
In the past 10 years, microprocessors have evolved at such a
rate that a modern chip can rival the computing power of a
mainframe computer. Such evolution has been possible
because of the increasing numbers of transistors that processors integrate. Pentium CPUs, for example, integrate well
over 5 million transistors on a single piece of silicon.
To integrate so many transistors on a piece of silicon, their
physical geometry has been reduced to the sub-micron level.
As a result of each geometry reduction, the corresponding
operational voltage for each transistor has also been reduced.
This changing CPU voltage demands the design of a programmable power supply–a design that is not completely
re-engineered with every change in CPU voltage.
The voltage range of the CPU has shown a downwards trend
for the past 5 years: from 3.3V for the Pentium, to 3.1V for
the Pentium Pro, to 2.8V for the Klamath, and to 2.0V for
the Deschutes processors. With this trend in mind, Fairchild
Semiconductor has designed the RC5051 controller. This
controller integrates the necessary programmability to
address the changing power supply requirements of lower
voltage CPUs.
Previous generations of DC-DC converter controllers were
designed with fixed output voltages adjustable only by
means of a set of external resistors. In a high volume production environment (such as with personal computers), however, a CPU voltage change would require a CPU board redesign to accommodate the new voltage requirement. The
5-bit DAC in the RC5051 reads the voltage ID code that is
programmed into modern processors and provides the appropriate CPU voltage. In this manner, the PC board does not
have to be re-designed each time the CPU voltage changes.
The CPU can thus automatically configure its own required
supply voltage.
Pentium is a registered trademark of Intel Corporation.
Intel Pentium II Processor Power
Requirements
Refer to Intel’s AP-587 Application Note, Slot 1 Processor
Power Distribution Guidelines, May 1997 (order number
243332-001), as a basic reference.
Input Voltages
Available inputs are +12V ±5% and +5V ±5%. Either one or
both of these inputs can be used by the DC-DC converter.
The input voltage requirements for Fairchild’s RC5051 DCDC converter are listed in Table 1. See below for detailed
information on how to apply these.
Table 1. Input Voltage Requirements
Part #
Vcc for IC
MOSFET
Drain
RC5051
+5V ±5%
+5V ±5% or 12V ±5%
Pentium II DC Power Requirements
Refer to Table 2, Intel Pentium II Processor Power Specifications. For standard motherboard designs , the on-board DC-DC
converter must supply a minimum of 14.2A at 2.8V for Klamath
or 2.0V for Deschutes. For a Deschutes Flexible Motherboard
design, the on-board DC-DC converter must supply 18.9A.
DC Voltage Regulation
As indicated in Table 2, the voltage level supplied to the
CPU must be within ±3% of its nominal setting. Voltage regulation limits must include:
•
•
•
•
Output load ranges specified in Table 2
Output ripple/noise
DC output initial voltage set point
Temperature and warm up drift (Ambient +0°C to +70°C
at full load with a maximum rate of change of 5°C per 10
minutes but no more than 10°C per hour)
• Output load transient with:
• Slew rate >30A/ms at converter pins Range:
0.3A - ICCP Max (as defined in Table 2).
Rev. 1.0.0
AN53
APPLICATION NOTE
Table 2. Intel Pentium II® Processor Power Specifications
Voltage Specification,
VCC CORE (VDC)
CPU Model, Features
Klamath
233MHz
266MHz
300MHz
Maximum Current, ICC CORE (A)
2.8V +100mV/-60mV
11.8
12.7
14.2
Deschutes
266MHz
300MHz
333MHz
350MHz
400MHz
450MHz
500MHz
2.0V +100mV/-60mV
2.0V +100mV/-60mV
2.0V +100mV/-60mV
2.0V +/-60mV
2.0V +/-60mV
2.0V +/-60mV
2.0V +/-60mV
8.5
9.6
10.6
11.1
12.1
14.2
16.0
Deschutes Flexible Motherboard
2.0V +/-60mV
18.9
NOTES:
1. Maximum power values are measured at typical VCCP to take into account the thermal time constant of the CPU package.
2. Flexible motherboard specifications are recommendations only. Actual specifications are subject to change.
Output Ripple and Noise
Ripple and noise are defined as periodic or random signals
over the frequency band of 0–20MHz at the output pins.
Output ripple and noise must be consistent with voltage
requirements throughout the full load range and under all
specified input voltage conditions.
Efficiency
The efficiency of the DC-DC converter must be greater than
80% at maximum output current and greater than 40% at low
current draw.
Processor Voltage Identification
There are five Voltage Identification Pins, VID4-VID0, on the
Pentium II processor package which can be used to support
automatic selection of the power supply voltage. These pins are
internally unconnected or are shorted to ground (VSS ). The
logic status of the VID pins defines the voltage required by the
processor. In order to address future low voltage microprocessors, the RC5051 includes a VID4 input bit to extend the output
voltage range as low as 1.3V. The output voltage programming
codes are presented in Table 3. A ‘1’ refers to an open pin and a
‘0’ refers to a short to ground.
Table 3. Output Voltage Programming
Codes
VID4
0
0
0
0
0
0
0
2
VID3
1
1
1
1
1
1
1
VID2
1
1
1
1
0
0
0
VID1
1
1
0
0
1
1
0
VID0 VOUT to CPU
1
1.30V
0
1.35V
1
1.40V
0
1.45V
1
1.50V
0
1.55V
1
1.60V
Table 3. Output Voltage Programming
Codes (continued)
VID4
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
VID3
1
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
VID2
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
VID1
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
Note:
1. 0 = processor pin is tied to GND
1 = processor pin is open.
VID0 VOUT to CPU
0
1.65V
1
1.70V
0
1.75V
1
1.80V
0
1.85V
1
1.90V
0
1.95V
1
2.00V
0
2.05V
1
No CPU
0
2.1V
1
2.2V
0
2.3V
1
2.4V
0
2.5V
1
2.6V
0
2.7V
1
2.8V
0
2.9V
1
3.0V
0
3.1V
1
3.2V
0
3.3V
1
3.4V
0
3.5V
APPLICATION NOTE
AN53
I/O Controls
The RC5051 Controller
In addition to the Voltage Identification, there are several signals that control the DC-DC converter or provide feedback from
the DC-DC converter to the CPU. They are Power-Good
(PWRGD), and Output Enable (OUTEN). These signals will be
discussed later.
The RC5051 is a programmable synchronous DC-DC
controller IC. When designed around the appropriate external components, this device can be configured to deliver
more than 14.5A of output current. The RC5051 utilizes both
current-mode and voltage-mode PWM control to create an
integrated step-down voltage regulator.
RC5051 Description
Main Control Loop
Simple Step-Down Converter
S1
L1
+
VIN
D1
C1
RL Vout
–
65-AP53-03
Figure 1. Simple Buck DC-DC Converter
For the purpose of understanding a buck converter, Figure 1
illustrates a step-down DC-DC converter with no feedback control. The operation of the basic step-down converter is the basis
for the design equations for the RC5051. Referring to Figure 1,
the basic operation begins by closing the switch S1. When S1 is
closed, the input voltage VIN is applied to inductor L1. The current flowing in this inductor increases, and the increase is given
by the following equation:
( V IN – V OUT )T ON
DI L = ---------------------------------------------L1
where TON is the time S1 is closed (the duty cycle is TON / TS,
with TS the switching period). When S1 opens, the diode D1
conducts the inductor current and the output current is delivered
to the load; the inductor current decrease is given by:
V OUT ( T S – T ON )
DI L = -----------------------------------------L1
where (TS - TON ) is the time during which S1 is open.
By solving these two equations, we can arrive at the basic relationship for the output voltage of a step-down converter:
T ON
V OUT = V IN ----------TS
Refer to the RC5051 Block Diagram illustrated in Figure 2.
The control loop of the regulator contains two main sections:
the analog control block and the digital control block. The
analog section consists of signal conditioning amplifiers
feeding into a set of comparators which provide the inputs to
the digital control block. The signal conditioning section
accepts inputs from the IFB (current feedback) and VFB
(voltage feedback) pins and sets up two controlling signal
paths. The voltage control path amplifies the difference
between the VFB signal and the voltage reference and presents the output to one of the summing amplifier inputs. The
current control path takes the difference between the IFB and
VFB pins and presents the resulting signal to another input
of the summing amplifier. These two signals are then
summed together with the slope compensation input from
the oscillator. This output is then presented to a comparator,
which provides the main PWM control signal to the digital
control block.
The additional comparators in the analog control section set
the point at which the current limit comparator disables the
output drive signals to the external power MOSFETs.
The digital control block takes the comparator inputs and the
main clock signal from the oscillator to provide the appropriate pulses to the HIDRV and LODRV output pins. These
pins control the external power MOSFETs. The digital section utilizes high speed Schottky transistor logic, allowing
the RC5051 to operate at clock speeds as high as 1MHz.
High Current Output Drivers
The RC5051 contains two identical high current output
drivers that utilize high speed bipolar transistors in a pushpull configuration. Each driver is capable of delivering 1A of
current in less than 100ns. Each driver’s power and ground
are separated from the chip’s power and ground for additional switching noise immunity.
In order to obtain a more accurate approximation for VOUT , we
must also include the forward voltage VD across diode D1 and
the voltage across the switch , VSW . After taking into account
these factors, the new relationship becomes:
T ON
V OUT = ( V IN + V D – V SW ) ----------- – V D
TS
where VSW = is the voltage across the MOSFET in the on state,
I L * R DS,ON.
3
AN53
APPLICATION NOTE
+12V
+5V
RC5051
1
5
4
–
+
OSC
13
–
+
–
+
–
+
12
+5V
DIGITAL
CONTROL
VO
7
9
VREF
16
1.24V
REFERENCE
5-BIT
DAC
20 19 18 17
8
POWER
GOOD
3
PWRGD
2
65-5051-01
ENABLE
VID0 VID2 VID4
VID1 VID3
Figure 2. RC5051 Block Diagram
The HIDRV driver has a power supply, VCCQP, supplied from a
12V source as illustrated in Figure 2. The resulting voltage is
sufficient to provide the gate to source voltage to the external
MOSFET that is required to achieve a low R DS,ON . Since the
low side synchronous FET is referenced to ground, there is no
need to boost the gate drive voltage, and its VCCP power pin can
be tied to VCC.
Output Enable (OUTEN)
Internal Voltage Reference
The RC5051 constantly monitors the output voltage for protection against over voltage conditions. If the voltage at the
VFB pin exceeds 20% of the selected program voltage, an
over-voltage condition is assumed and the chip disables the
output drive signal to the external MOSFETs.
The reference included in the RC5051 is a precision bandgap voltage reference. The internal resistors are precisely
trimmed to provide a near zero temperature coefficient (TC).
Added to the reference input is the resulting output from an
integrated 5-bit DAC–provided in accordance with the Pentium II specification guidelines. These guidelines require the
DC-DC converter output to be directly programmable via a
5-bit voltage identification (VID) code. This code scales the
reference voltage from 2.0V (no CPU) to 3.5V in 100mV
increments, and between 1.3V and 2.05V in 50mV increments. For guaranteed stable operation under all operating
conditions, 0.1mF of decoupling capacitance should be
connected to the VREF pin. No load should be imposed on
this pin.
Power Good (PWRGD)
The RC5051 Power Good function is designed in accordance
with the Pentium II DC-DC converter specification to provide a constant voltage monitor on the VFB pin. The circuit
compares the VFB signal to the VREF voltage and outputs an
active-low interrupt signal to the CPU when the power supply voltage differs more than ±12% from nominal. The
Power Good flag provides no other control function to the
RC5051.
4
The DC-DC converter accepts an open collector signal for
controlling the output voltage. The low state disables the
output voltage. When disabled, the PWRGD output is in the
low state.
Over-Voltage Protection
Over-current Protection
A current sense methodology is implemented to disable the
output drive signal to the MOSFETs when an over-current
condition is detected. The voltage drop created by the output
current flowing across a sense resistor is presented to an
internal comparator. When the voltage developed across the
sense resistor exceeds the comparator threshold voltage, the
chip reduces the output drive signal to the MOSFETs.
The DC-DC converter returns to normal operation after the
fault has been removed, for either an over-voltage or an overcurrent condition.
Oscillator
The RC5051 oscillator section uses a fixed current capacitor
charging configuration. An external capacitor (C EXT ) is used to
preset the oscillator frequency between 80KHz and 1MHz.
This scheme allows maximum flexibility in choosing external
components.
APPLICATION NOTE
AN53
In general, a higher operating frequency decreases the peak
ripple current flowing in the output inductor, thus allowing
the use of a smaller inductor value. operation at higher frequencies also decreases the amount of energy storage that
must be provided by the bulk output capacitors during load
transients.
An operating frequency of 300 kHz was chosen in this
Application Note to optimize efficiency while maintaining
excellent regulation and transient performance under all
operating conditions.
Design Considerations and Component
Selection
Unfortunately, the efficiency losses due to switching of the
MOSFETs increase as the operating frequency is increased.
Thus, efficiency is optimized at lower operating frequencies.
Figure 3 illustrates a synchronous application using the
RC5051.
+12V
L1
+5V
C1
2.5mH
CIN*
0.1mF
C2
0.1mF
R1
47½
D1
1N4735A
C6
0.1mF
C5
1mF
R2
Q1
VREF
C3
0.1mF
11
10
9
12
8
13
7
14
15
6
RC5051
16
5
17
4
18
3
19
2
20
1
C4
1mF
4.7½
L2
R3
2.3µH
Q2
4.7½
R SENSE*
VO
COUT*
D2
1N5820
CEXT
100pF
VID4
VID3
VID2
VCC
ENABLE
C7
0.1mF
R4
10K½
C8
0.1mF
PWRGD
*Refer to Table 3 for values of
RSENSE, COUT, and CIN.
VID1
VID0
65-5051-03
Figure 3. Synchronous DC-DC Converter Application Schematic Using the RC5051
5
AN53
APPLICATION NOTE
Table 4. RC5051 Application Bill of Materials for Intel Pentium II Processors
Reference
Manufacturer’s Part #
Quantity
Description
Requirements/Comments
C1-3, C6-8
6
100nF, 50V Capacitor
C4-5
2
1mF 16V Capacitor
Cext
1
100pF Capacitor
5%, C0G
CIN
Sanyo
10MV1200GX
*
1200mF, 10V electrolytic
IRMS = 2A
*See Table 5
COUT
Sanyo
6MV1500GX
*
1500mF, 6.3V electrolytic
ESR < 44mW
*See Table 5
D1
Motorola
1N4735A
1
6.2V Zener Diode
5%
D2
Motorola
IN5820
1
3A Schottky Diode
1
2.5mH Inductor
DCR < 6mW1
L1
L2
Elytone
YT-6542
1
1.3mH Inductor
DCR < 3mW
Q1-2
Fairchild
FDP6030L or FDB6030L
2
N-channel MOSFET
RDS, ON = 20mW @
VGS = 4.5V2
1
47W
1/10W
R1
R2-3
2
4.7W
1/10W
RSENSE
Fairchild
RC10-XX
1
CuNi Alloy Wire Resistor
*See Table 5
U1
Fairchild
RC5051M
1
DC/DC Controller
Notes:
1. Inductor L1 is recommended to isolate the 5V input supply from noise generated by the MOSFET switching, and to comply
with Intel dI/dt requirements. If used, it must be sized to avoid saturation at maximum input current. L1 may be omitted if
desired.
2. For 14.2A designs using the FDP6030L MOSFETs, heatsinks with thermal resistance QSA <20°C/W should be used. For details and a spreadsheet on MOSFET selections, refer to Applications Bulletin AB-8.
Table 5. Recommended Values for CPU-based Applications
Application
300MHz AMD K6
Motherboard
6
Output
Current
13A
CIN
COUT
COUT Maximum
ESR
RSENSE
3 x 1200mF, 10V
Sanyo
10MV1200GX
7 x 1500mF, 6.3V
Sanyo
6MV1500GX
6.1mW
5.8mW
300MHz Intel Pentium II 14.2A
Klamath Motherboard
3 x 1200mF, 10V
Sanyo
10MV1200GX
7 x 1500mF, 6.3V
Sanyo
6MV1500GX
6.8mW
5.2mW
300MHz Intel Pentium II 12.6A
Deschutes Motherboard
3 x 1200mF, 10V
Sanyo
10MV1200GX
7 x 1500mF, 6.3V
Sanyo
6MV1500GX
5.3mW
5.8mW
APPLICATION NOTE
AN53
MOSFET Selection Cosiderations
MOSFET Selection
This application requires N-channel Logic Level Enhancement Mode Field Effect Transistors. Desired characteristics
are as follows:
The on-resistance (RDS,ON ) is the primary parameter for
MOSFET selection. It determines the power dissipation
within the MOSFET and, therefore, significantly affects the
efficiency of the DC-DC converter. Table 6 is a selection
table for MOSFETs.
• Low Static Drain-Source On-Resistance, RDS,ON < 40 mW
(lower is better)
• Low gate drive voltage, VGS £ 4.5V
• Power package with low Thermal Resistance
• Drain current rating of 20A minimum
• Drain-Source voltage > 15V.
Table 6. MOSFET Selection Table
Manufacturer
and Part #
RDS, ON (mW)
Conditions
Typ.
Max.
Package
Thermal
Resistance (°C/W)
Fairchild
FDP6030L
VGS = 4.5V, ID = 21A
TJ = 25°C
15
20
TO-220
QJA = 62.5
Fairchild
FDB6030L
VGS = 4.5V, ID = 21A
TJ = 25°C
15
20
D2-PAK
QJA = 62.5
Fairchild
FDP603AL
VGS = 4.5V, ID = 10A
TJ = 25°C
30
36
TO-220
QJA = 62.5
Fairchild
FDB603AL
VGS = 4.5V, ID = 10A
TJ = 25°C
30
36
D2-PAK
QJA = 62.5
Fairchild
FDP7030L
VGS = 5V, ID = 40A
TJ = 25°C
9
10
TO-220
QJA = 62.5
Fairchild
FDB7030L
VGS = 5V, ID = 40A
TJ = 25°C
9
10
D2-PAK
QJA = 62.5
IR
IRL2203N
VGS = 4.5V, ID = 50A
TJ = 25°C
10
TO-220
QJA = 62
IR
IRL2203S
VGS = 4.5V, ID = 50A
TJ = 25°C
10
D2-PAK0
QJA = 40
7
AN53
APPLICATION NOTE
Two MOSFETs in parallel.
+5V
If output current is high, We recommend two MOSFETs
used in parallel instead of one single MOSFET. The following significant advantages are realized using two MOSFETs
in parallel:
D1
VCCQP
Q1
HIDRV
• Significant reduction of Power dissipation.
Example: RC5051 with Maximum output current of 14A
at 2.8V with one MOSFET on the high side:
CP
L2
RS
VO
PWM/PFM
Control
COUT
LODRV
PMOSFET = (I2 RDS,ON)(Duty Cycle) =
(14A)2 (0.050W)(2.8V / 5V) = 5.5W
Q2
D2
GNDP
With two MOSFETs in parallel:
65-5051-06
Figure 4. Charge Pump Configuration
2R
PMOSFET = (I
DS,ON )(Duty Cycle) =
(14A/2)2 (0.037W) (2.8V / 5V) = 1.0W/FET
*Note: RDS,ON increases with temperature. Assume RDS,ON =
25mW at 25°C. RDS, ON can easily increase to 50mW at high temperature when using a single MOSFET. When using two MOSFETs in parallel, the temperature effects should not cause the RDS, ON to rise as
much.
• Smaller heat sink required.
With power dissipation down to around one watt,
considerably less heat sink is required.
• Reliability.
With thermal management under control, this DC-DC
converter is able to deliver load currents up to 14.5A with
no performance or reliability concerns.
• Method 2. 12V Gate Bias.
Figure 5 illustrates how a 12V source can be used to bias the
VCCQP. A 47W resistor is used to limit the transient current
into the VCCQP pin and a 1mF capacitor filter is used to filter
the VCCQP supply. This method provides a higher gate bias
voltage (VGS ) to the MOSFET than the charge-pump method
does, and therefore reduces the RDS, ON of the MOSFET and
thus reduces the power loss due to the MOSFET. Figure 6
shows how RDS, ON reduces dramatically with VGS increases.
A 6.2V Zener diode (D1) is placed from VCCQP to 5V to
clamp the voltage at VCCQP to a maximum of 12V and
ensure that the absolute maximum voltage of the IC will not
be exceeded.
+5V
• MOSFET Gate Bias.
As already discussed, the low-side MOSFET on the
RC5051 needs only 5V for its gate drive supply. The highside MOSFET can be biased by one of two methods:
Charge Pump or 12V Gate Bias.
47½
D1
VCCQP
Q1
HIDRV
1µF
L2
RS
VO
PWM/PFM
Control
COUT
LODRV
Q2
D2
GNDP
65-5051-07
Figure 5. 12V Gate Bias Configuration
Rds, on (W)
• Method 1. Charge pump (or Bootstrap) method.
Figure 4 displays the use of a charge pump to provide gate
bias to the high-side MOSFET with the RC5051.
Capacitor CP is the charge pump deployed to boost the
voltage of the RC5051 output driver. When the MOSFET
switches off, the source of the MOSFET is at
approximately 0V. VCCQP is charged through the
Schottky diode D1 to approximately 4.5V. Thus, the
capacitor CP is charged to approximately 4.5V. When the
MOSFET turns on, the source of the MOSFET voltage is
equal to 5V. The capacitor voltage follows, and hence
provides a voltage at VCCQP equal to approximately 10V.
The Schottky diode D1 is required to provide the charge
path when the MOSFET is off, and reverses bias when the
VCCQP goes to 10V. The charge pump capacitor, CP,
needs to be a high Q, high frequency capacitor. A 1mF
ceramic capacitor is recommended here.
+12V
IRL2203NS
NDP6030L
FDP7030L
NDB603AL
Figure 6. RDS,ON vs. VGS for Selected MOSFETs
8
APPLICATION NOTE
AN53
Converter Efficiency
Losses due to parasitic resistance in the switches, coil, and
sense resistor dominate at high load-current level. The major
loss mechanisms under heavy loads, in typical order of
importance, are:
•
•
•
•
•
•
Sense Resistor losses
Input Capacitor losses
Coil Losses
Losses due to the operating supply current of the IC.
Diode-conduction losses
Gate-charge losses
• MOSFET I 2 R losses
• Transition losses
Formulae for Calculation of Converter Efficiency
P OUT
V OUT I OUT
= ------------------------------------------------Efficiency = ------------p IN
V OUT I OUT + P LOSS
P LOSS = P RDS + ( P RISE + P FALL ) + P INDUCTOR + P SENSE + P GATE + P DIODE + P CAPS + P IC
2
P RDS = I OUT R DS,ON DC for the high-side MOSFET
2
P RDS = I OUT R DS,ON ( 1 – DC ) for the low-side MOSFET
V IN I OUT t RISE F S V IN I OUT t FALL F S
P RISE + P FALL = ----------------------------------------- + -------------------------------------------- for the high-side MOSFET
2
2
V f I OUT t RISE F S V f I OUT t FALL F S
---------------------------------------P RISE + P FALL = --------------------------------------- +
for the low-side MOSFET
2
2
2
P INDUCTOR = I OUT R INDUCTOR
2
P SENSE = I OUT R SENSE
2
P GATE = CV F S for each MOSFET
P DIODE = I OUT V F T DT F S , with TDT the deadtime
2
P CAPS = ESR ´ I OUT ´ DC ´ ( 1 – DC )
P IC = 25mA* V CC
Example Efficiency Calculation
As an example, the efficiency of a synchronous 14A converter based on the RC5051 will be calculated. The converter produces
2.8V output from a 5V input, switches at 300kHz, has MOSFETs with a 4nF gate capacitance, RDS,on = 10mW for both MOSFETS and a rise and fall time of 50nsec. The inductor has a winding resistance of 3mW, and the sense resistor used is 5.2mW. The
schottky paralleled with the synchronous rectifier has a forward voltage of 400mV at 14A. The input capacitors have a total ESR
of 15mW.
2
P RDS = ( 14A ) 10mW = 1.96W
5V ´ 14A ( 50nsec + 50nsec )300kHz
P RISE + P FALL = ------------------------------------------------------------------------------------------- = 1.05W
2
400mV ´ 14A ( 50nsec + 50nsec )300kHz
P RISE + P FALL = ------------------------------------------------------------------------------------------------------ = 84mW
2
2
P INDUCTOR = ( 14A ) 3mW = 590mW
2
P SENSE = ( 14A ) 5.2mW = 1.02W
2
P GATE = 4nF ( 5V ) 300kHz ´ 2 = 60mW
P DIODE = 14A ´ 400mV ´ 50nsec ´ 300kHz = 84mW
2
P CAPS = 15mW ´ ( 14A ) ´ 0.56 ´ ( 1 – 0.56 ) = 724mW
P IC = 25mA ´ 5V ´ 125mW
P LOSS = 1.96W + 1.05W + 0.084W + 0.59W + 1.02W + 0.06W + 0.084W + 0.724W + 0.125W = 5.70W
2.8V ´ 14A
Efficiency = ---------------------------------------------------- = 87%
2.8V ´ 14A + 5.70W
9
AN53
APPLICATION NOTE
When using these formulae, special care must be taken
regarding the MOSFETs’ transition times: the rise and fall
refer to the MOSFETs’ drain-source voltage, NOT the gatesource. Using the datasheet values (rather than measured values) can also result in serious overestimation of the losses,
since the transition is being driven by an inductive source,
not a resistor.
Selecting the Inductor
The inductor is one of the most critical components to be
selected for a DC-DC converter application. The critical parameters of the inductor are its inductance (L), maximum DC current (IO ), and DC coil resistance (Rl ).
The inductor’s inductance helps determine two key parameters of a converter, its ripple current and its transient
response. On the one hand, making the inductance large
reduces the ripple current, and thus the output ripple voltage.
On the other hand, a large inductance provides a slow
response to load transients. For Pentium II supplies, the transient response is paramount, and thus the inductance is typically chosen to be in the 1-5mH range.
When designing the external current sense circuitry, pay
careful attention to the output limitations during normal
operation and during a fault condition. If the over-current
protection threshold current is set too low, the DC-DC converter may not be able to continuously deliver the maximum
CPU load current. If the threshold level is too high, the output driver may not be disabled at a safe limit and the resulting power dissipation within the MOSFETs may rise to
destructive levels. The following is the design equation used
to set the over-current threshold limit:
I RIPPLE
I PK = I LOAD, MAX + -----------------2
Where I pk is defined as in Figure 7, and Iload, max = maximum
output load current. Figure 7 illustrates the inductor current
waveform for the RC5051 DC-DC converter at maximum load.
IPK
I
ILOAD, MAX
IMIN
TON
Most inductors’ inductance also depends on current, that is,
increasing the current through the inductor decreases the
inductance. It is thus vital to specify the DC current when
procuring an inductor. The one type of inductor which does
not change inductance with current is the rod-core inductor,
but this type may have significant EMI (noise) problems. For
further information, refer to Applications Bulletin AB-12.
The resistance of the winding of the inductor is also important, as it is directly responsible for much of the losses in the
inductor. Minimizing the resistance will help improve the
converter’s efficiency.
t
TOFF
Ts
65-AP53-04
Figure 7. Typical DC-DC Converter Inductor
Current Waveform
The calculation of the ripple current is as follows:
V IN – V OUT V OUT
I RIPPLE
------------------ = ------------------------------T S -------------2L
V IN
2
where:
VIN = input voltage to converter,
Implementing Over-current Protection
TS = the switching period of the converter = 1/fS, and
Intel currently requires all power supply manufacturers to
provide continuous protection against short circuit conditions that may damage the CPU. To address this requirement,
Fairchild Semiconductor has implemented a current sense
methodology to limit the power delivered to the load in the
event of over-current. The voltage drop created by the output
current across a sense resistor is presented to one terminal of
an internal comparator with hysteresis. The other comparator
terminal has the threshold voltage, nominally of 120mV.
Table 7 states the limits for the comparator threshold of the
Switching Regulator.
fS = switching frequency.
Table 7. RC5051 Over-current Comparator
Threshold Voltage
Short Circuit Comparator
Vthreshold (mV)
Typical
120
Minimum
100
Maximum
140
10
As an example, for an input voltage of 5V, output voltage of
2.8V @ 14A, L equal to 1.3mH and a switching frequency of
285KHz (using CEXT = 100pF), the peak inductor current can
be calculated as :
5V – 2.8V
1
2.8V
I PK = 14A + ------------------------- ´ -------------------- ´ ------------ = 15.7A
2 ´ 1.3mH 285kHz 5V
Therefore, the over-current detection threshold must be at
least 16A. The next step is to determine the value of the
sense resistor. Including sense resistor tolerance, the sense
resistor value can be determined as
V th,min
R SENSE = ----------------------------I SC ( 1 + TF )
Where TF = Tolerance Factor for the sense resistor. Table 8
describes tolerance, size, power capability, temperature coefficient and cost of various type of sense resistors.
AN53
APPLICATION NOTE
Table 8. Comparison of Sense Resistors
Discrete Iron
Alloy
Resistor (IRC)
Motherboard
Trace Resistor
Description
Discrete Metal
Strip Surface
Mount Resistor
(Dale)
Discrete MnCu
Alloy Wire
Resistor
Discrete
CuNi Alloy
Wire Resistor
(Copel)
Tolerance Factor
(TF)
±20%
±5%
(±1% available)
±1%
±10%
±10%
Size (L x W x H)
2" x 0.2" x 0.001"
(1 oz Cu trace)
0.45" x 0.065" x
0.200"
0.25" x 0.125" x
0.025"
0.200" x 0.04" x
0.160"
0.200" x 0.04" x
0.100"
Power capability
>50A/in
1 watt
(3W and 5W
available)
1 watt
(2W available)
>1 watt
>1 watt
+3900 ppm
+390 ppm
±75 ppm
±50 ppm
±20 ppm
Low---included in
motherboard
$0.31
$0.47
$0.09
$0.09
Temperature
Coefficient
Cost @10,000
piece
Continuing with the example, based on the Tolerance Factor
in the above table, for an embedded PC trace resistor and for
ISC set 1A greater than Ipk:
100mV
R SENSE = ----------------------------------------------------------- = 5.0mW
( 15.7A + 1A ) ( 1 + 0.20 )
Comment on Discrete Sense Resistors
Discrete Iron Alloy resistors come in a variety of tolerances
and power ratings, and are ideal for precision implementation, minimizing stress on the MOSFETs. MnCu Alloy wire
resistors or CuNi Alloy wire resistors are ideal for low cost
implementations.
For a discrete Cuni resistor:
Designing an Embedded Sense Resistor (PC Trace
Resistor)
100mV
R SENSE = ----------------------------------------------------------- = 5.7mW
( 15.7A + 1A ) ( 1 + 0.10 )
For user convenience, Table 9 lists the recommended values
for sense resistors for various load currents using embedded
trace resistors and discrete resistors and assuming (Ipk- Imin) /
2 = 10% IOUT and ISC set 1A greater than Ipk
Table 9. Rsense for Various Load Currents
ILoad,max (A)
RSENSE PC Trace
Resistor (mW)
RSENSE Discrete
Resistor (mW)
6.9
9.7
10.6
7.8
8.7
9.5
8.5
8.1
8.8
8.7
7.9
8.6
9.6
7.2
7.9
10.6
6.6
7.2
11.1
6.3
6.9
12.6
5.6
6.1
14.2
5.0
5.5
17.2
4.2
4.6
18.5
3.9
4.3
18.9
3.8
4.2
11
Embedded PC trace resistors have the advantage of near zero
cost implementation. However, the value of the PC trace
resistance has large variations. To start with, traces on the
outside layers have far more variation than those on inside
layers; for this reason, embedded resistors should always be
designed on inside layers. Embedded resistors have 3 error
sources: the sheet resistivity of the layer, the tolerance of the
width and length of the trace, and the temperature variation
of the copper. Only two of the error sources must be considered for laying out embedded sense resistors.
• Sheet resistivity.
For 1 ounce copper, the thickness is nominally 1.3mils,
and the toleranace is typically ±.1 mil. Therefore error due
to sheet resistivity is 0.1 mil / 1.3mil = +/-8%
• Tolerance of width and length.
The width and length of traces also have tolerances;
typical numbers might be ±3/4 mil. A typical width to
obtain a 5.2mW resistor might be 300mils, and so the
length would have to be:
t´w´R
1.3mil ´ 300mil ´ 5.2mW
L = ---------------------- = ---------------------------------------------------------------- = 2996mils
–4
r
6.77 ´ 10 W – mil
or 3 inches. The tolerance is thus negligible.
AN53
APPLICATION NOTE
• Thermal Consideration.
Due to I2R power losses the temperature of the resistor will
increase, leading to a higher resistance value. In addition,
ambient temperature variation will add to the change in
resistor value:
R = R 20C ´ 1.00393
Refer to Figure 8.
1 Each
W = 300 mils
L = 3000mils
T – 20
65-AP53-05
where: R20C is the resistance at 20°C, T is the trace’s
temperature, and R is the actual resistance. For example,
for temperature T = 50°C, the %R change = 12%. Table
10 is the summary of the tolerance for the Embedded PC
Trace Resistor.
Table 10. Summary PC Trace Resistor
Tolerance
Figure 8. 5.20mW Sense Resistor (10 ■ )
You can also implement the sense resistor in the following
manner. Each corner square is counted as 0.6 square since
current flowing through the corner square does not flow
uniformly and it is concentrated towards the inside edge,
as shown in Figure 9.
1 Each
Tolerance due to Sheet Resistivity variation
8%
Tolerance due to temperature variation
12%
Total Tolerance for PC Trace Resistor
20%
.6
.6
.8
Design Rules for Using an Embedded Resistor
The basic equation for the resistance of an embedded
resistor is:
L
R = r ´ ------------W´t
where:
r = Resistivity(mW-mil),
L = Length(mils),
W = Width(mils), and
t = Thickness(mils).
For 1oz copper, t = 1.3 mils, r = 677mW-mil,
1 L/1 W = 1 Square ( ■ ).
For example, you can layout a 5.20mW embedded sense
resistor using the example above
L/W = 300mils
RWt
0.0052 ´ 300 ´ 1.3
L = ------------ = --------------------------------------------- = 3000mils
r
677
65-AP53-06
Figure 9. 5.30mW Sense Resistor (10 ■)
A Design Example Combining an Embedded Resistor
and a Discrete Resistor
For low cost implementation, the embedded PC trace resistor
is most desirable. However, its wide tolerance (20%) can
present a problem. In addition, requirements for the CPU
change frequently, and, thus, the maximum load current may
be subject to change. Combining embedded resistors with
discrete resistors may be a desirable option. Figure 10 shows
a design that provides flexibility with a solution to address
wide tolerances. In this design, you have the option to
choose an embedded or a discrete sense resistor. To use the
discrete sense resistor, populate R21 with a shorting bar
(zero Ohm resistor) for proper Kelvin connection and add
the MnCu sense resistor. To use the embedded sense resistor,
on the other hand, populate R22 with a shorting bar for
Kelvin connection. The embedded sense resistor allows the
user to choose a plus or a minus delta resistance tap to offset
any large sheet resistivity change. In this design, the center
tap yields 6mW, the left tap yields 6.8mW, and the right tap
yields 5.3mW.
L/W = 10 ■
IFB
MnCu Discrete
VFB
Resistor
R21
R22
Output Power
Plane (Vout)
R-Dr
R
R+Dr
65-AP53-07
Figure 10. Short Circuit Sense Resistor Design Using a PC Trace Resistor and an Optional Discrete Sense Resistor
12
APPLICATION NOTE
AN53
RC5051 over-current Characteristics
The RC5051 over-current characteristic includes a hysteresis
function that prevents the DC-DC converter from oscillating
in the event of an over-current. Figure 11 shows the typical
characteristic of the DC-DC converter circuit with a 6mW
sense resistor. The converter exhibits a normal load regulation characteristic until the voltage across the resistor
exceeds the internal over-current threshold of 120mV. At this
point, the internal comparator trips and signals the controller
to reduce the duty cycle of the high-side MOSFET. This
causes a drastic reduction in output voltage as the load
regulation collapses into the over-current control mode.
The output voltage does not return to its nominal value until
the output current is reduced to a value within the safe range
for the DC-DC converter.
These calculations show that the high-side MOSFET is not
being over-stressed during an over-current condition.
Output Voltage vs. Output Current
RSENSE = 6m½
3.5
3.0
Figure 12A. HIDRV Output Waveform for Normal
Operation Condition with Vout = 2.8V@10A
OUT (V)
2.5
2.0
1.5
65-5051-08
1.0
0.5
0
0
5
10
15
20
25
Output Current (A)
Figure 11. RC5051 Over-current Characteristic
Power Dissipation Consideration During an
Over-current Condition
The RC5051 controller responds to an output over-current by
drastically reducing the duty cycle of the gate drive signal to
the high-side MOSFET. In doing this, the high-side MOSFET is protected from stress and from eventual failure. Figure 12A shows the gate drive signal of a typical RC5051
operating in continuous mode with a load current of 10A.
The duty cycle is set by the ratio of the input voltage to the
output voltage. If the input voltage is 5V, and the output voltage is 2.8V, the ratio of Vout/ Vin is 56% (64% measured).
Figure 12B shows the result of a RC5051 going into its overcurrent mode with a duty cycle of approximately 47%. Calculating the power in each MOSFET at each condition on the
graph (Figure 11) shows how the protection works. The
power dissipated in the high-side MOSFET at normal operation for a load current of 14.2A, is given by:
PD
=
2
2
I ´ R DS, ON ´ DC = ( 14.2 ) ´ .01
´ .64 =
1.29W
ignoring switching losses.
The power dissipated in the MOSFET at an over-current
condition of 20A, is given by:
2
P D = ( 20 ) ´ .01 ´ .47 = 1.88W
again ignoring switching losses.
Figure 12B. HIDRV Output Waveform for
Over-current Condition
Power dissipation in the low-side MOSFET during an overcurrent condition must also be considered. The low-side
MOSFET dissipates power while the high-side MOSFET is
off. The power dissipated in the low-side MOSFET during
normal operation, is given by:
=
2
2
I ´ R RDS, ON ´ ( 1 – DC ) = ( 14.2 ) ´ .01
´ .36 =
0.73
During an over-current , the duty cycle reduces to around
47%. The power dissipated in the low-side MOSFET during
short circuit condition, is given by:
2
P D = ( 20 ) ´ .01 ´ .53 = 2.1W
Thus, for the low-side MOSFET, the thermal dissipation during over-current is greatly magnified. This requires that the
thermal dissipation of the low-side MOSFET be properly
managed by an appropriate heat sink. To protect the low-side
MOSFET from being destroyed in the event of an over-cur13
AN53
APPLICATION NOTE
rent , you should limit the junction temperature to less than
130°C. You can find the required thermal resistance using the
equation for maximum junction temperature:
T J ( max ) – T A
P D = ------------------------------R QJA
the total ESR of the capacitors used and the parasitic resistance of the output traces. For a detailed analysis of capacitor
requirements in a high-end microprocessor system, please
refer to Application Bulletin 14.
Input Filter
Assuming that the ambient temperature is 50°C,
T J ( max ) – T A
130 – 50
R QJA = ------------------------------= --------------------- = 38°C ¤ W
PD
2.1
Thus, you need to provide a heat sink that gives the low-side
MOSFET a thermal resistance of 38°C/W or lower to protect
the device during an indefinite short.
In summary, with proper heat sink, the low-side MOSFET is
not over-stressed during an over-current condition.
Schottky Diode Selection
The application circuit diagram of Figure 3 shows a Schottky
diode, D2. D2 is used as a flyback diode to provide a current
path for the inductor current during the dead-time when both
the high-side and low-side MOSFETs are briefly both turned
off. Table 11 shows the characteristics of several Schottky
diodes. Note that MBRB2515L has a very low
forward voltage drop even at high current. Although it is not
necessary to use a high-current diode for this application,
selecting a higher current schottky will provide improved
efficiency at slightly higher cost.
The DC-DC converter may include an input inductor
between the system +5V supply and the converter input
as described below. This inductor serves to isolate the +5V
supply from the noise in the switching portion of the DC-DC
converter, and to limit the inrush current into the input capacitors during power up. A value of 2.5mH is typical, as illustrated in Figure 13; details on selection of an input inductor
may be found in Applications Bulletin AB-16.
The number of input capacitors required for a converter is
determined by the capacitors’ ripple current rating. The
ripple current is given by:
I rms = I OUT DC – DC
2
Thus, for example, a Deschutes processor running at 2.0V
out from 5.0V in has a DC = 2.0/5.0 = .4; if it pulls 14.2A, its
Irms = 7A.
Table 12 shows some typical input capacitors’ current ratings; the current rating increases as temperature decreases.
Although exceeding these ratings will not cause capacitor
damage, it will reduce their life, and thus the converter’s
MTBF.
Table 11. Schottky Diode Selection Table
Manufacturer
Model #
Conditions
Motorola
1N5817
IF = 1A; Tj = 25°C
<.45v
Motorola
1N5820
IF = 3A; Tj = 25°C
<.475v
Motorola
IF = 20A; Tj = 25°C
MBR2015CTL IF = 20A;Tj = 150°C
Motorola
IF = 19A; Tj = 70°C
MBRB2515L
Vin
< 0.28v
Correct calculation of the output capacitance is crucial to the
performance of the DC-DC converter. The output capacitor
determines the overall loop stability, output voltage ripple,
and, most importantly, load transient response. Because the
control loop response of the controller is not instantaneous,
the initial load transient must be supplied entirely by the output capacitors. The initial voltage deviation is determined by
1000µF, 10V
Electrolytic
0.1µF
65-5051-09
< 0.58v
< 0.48v
Output Filter Capacitors
14
5V
Forward
Voltage VF
Figure 13. Typical Input Filter
Table 12. Input Capacitor Selection Guide
Manufacturer
Part #
Irms
2.0A @
65°C
Sanyo
10MV1200GX
United
Chemicon
LXZ10VB122M10X20
1.2A @
105°C
Panasonic
EEUFA10122
1.2A @
105°C
AN53
APPLICATION NOTE
Q1
Q2
Good layout
11
RC505
Problem layout
10
12
9
13
8
14
7
15
6
16
5
17
4
18
3
19
2
20
1
Q2
= “Quiet" Pins
Q1
65-AP53-08
Figure 14. Placement of the MOSFETs
PCB Layout Guidelines and
Considerations
PCB Layout Guidelines
• Placement of the MOSFETs relative to the RC5051 is
critical. Place the MOSFETs (Q1 & Q2) so that the trace
length from the RC5051 HIDRV and LODRV pins to the
FET gates is minimized. A long lead length on this pin
would cause high amounts of ringing due to the
inductance of the trace and the large gate capacitance of
the FET. This noise radiates all throughout the board, and,
because it is switching at such a high voltage and
frequency, it is very difficult to suppress.
Figure 14 shows an example of good placement for the
MOSFETs in relation to the RC5051. In addition, this
figure shows an example of problematic placement for the
MOSFETs.
In general, all of the noisy switching lines should be kept
away from the quiet analog section of the RC5051. That is,
traces that connect to pins 9, 12 and 13 (LODRV, HIDRV
and VCCQP) should be kept far away from the traces that
connect to pins 1, 2, 4 and 5, and pin 16.
• Place the 0.1mF decoupling capacitors as close to the
RC5051 pins as possible. Extra lead length degrades their
ability to suppress noise.
• Each VCC and GND pin should have its own via to the
appropriate plane. This helps to provide isolation between
pins.
• Surround the CEXT timing capacitor with a ground trace.
Be sure to place a ground or power plane under the
15
capacitor for further noise isolation to provide additional
shielding to the oscillator pin 1 from the noise on the PCB.
In addition, place this capacitor as close to the RC5051
pin 1 as possible.
• Place the MOSFETs, inductor and Schottky as close
together as possible for the same reasons as on the first
bullet above. Place the input bulk capacitors as close to the
drains of MOSFETs as possible. In addition, placement of
a 0.1mF decoupling capacitor right on the drain of each
MOSFET helps to suppress some of the high frequency
switching noise on the input of the DC-DC converter.
• Place the output bulk capacitors as close to the CPU as
possible to optimize their ability to supply instantaneous
current to the load in the event of a current transient.
Additional space between the output capacitors and the
CPU allows the parasitic resistance of the board traces to
degrade the DC-DC converter’s performance under load
transient conditions, causing higher voltage deviation. For
more detailed information regarding capacitor placement,
refer to Application Bulletin AB-5.
• The traces that run from the RC5051 IFB (pin 4) and VFB
(pin 5) pins should be run next to each other and Kelvin
connected to the sense resistor. Running these lines
together prevents some of the common mode noise that is
presented to the RC5051 feedback input. Try, as much as
possible, to run the noisy switching signals (LODRV,
HIDRV & VCCQP) on one layer, but use the inner layers
for power and ground only. If the top layer is being used to
route all of the noisy switching signals, use the bottom
layer to route the analog sensing signals VFB and IFB.
AN53
Example of a PC Motherboard Layout and
Gerber File.
This section shows a reference design for motherboard
implementation of the RC5051 along with the Layout Gerber
File and Silk Screen. The actual PCAD Gerber File can be
obtained from Fairchild Semiconductor local Sales Office .
APPLICATION NOTE
7.
Next check the oscillator pin. You should see a saw tooth
wave at the frequency set by the external capacitor.
8.
When the VREF and CEXT pins are checked and
correct and the output voltage is incorrect, look at the
waveform at VCCQP. This pin should be +12V (in the
+12V application), and should be swinging from
slightly below +5V to about +10V (in the charge pump
application). If the VCCQP pin is noisy, with ripples/
over-shoots riding on it this may make the converter
function incorrectly.
9.
Next, look at HIDRV pin. This pin directly drives the
gate of the high-side FET. It should provide a gate drive
(measured gate to ground) of about 10V when turning
the FET on. A careful study of the layout is recommended. Refer to the “PCB Layout Guidelines” section.
Guidelines for Debugging and
Performance Evaluations
Debugging Your First Design Implementation
1.
Note the setting of the VID pins to know what voltage is
to be expected.
2.
Do not connect any load to the circuit. While monitoring
the output voltage, apply power to the circuit with current limiting at the power source. This ensures that no
catastrophic shorts are present.
3.
If proper voltage is not achieved go to “Procedures”
below.
4.
When you have proper voltage, increase the current
limiting of the power source to 16A.
5.
Apply load in 1A increments. An active load (HP6060B
or equivalent) is suggested.
6.
In case of poor regulation refer to “Procedures” below.
Procedures
1.
2.
3.
4.
5.
6.
16
10. Past experience shows that the most frequent errors are
incorrect components, improper connections, and poor
layout.
Performance Evaluation
This section shows sample evaluation results as a reference
guide for evaluating a DC-DC Converter using the RC5051
on a Pentium II motherboard.
DC Regulation
VID(2.0V)
43210
Iload (A)
Vout (V)
2.0V 00001
0.5
2.019
If there is no voltage at the output and the circuit is not
drawing current look for opens in the connections,
check the circuitry versus schematic, and check the
power supply pins of the RC5051 to make sure that voltage(s) are applied.
1
2.018
2
2.017
3
2.016
4
2.014
If there is no voltage at the output and the circuit is
drawing excessive current (>100mA) with no load,
check for possible shorts. Determine the path of the
excessive current and which device is drawing it–this
current may be drawn by peripheral components.
5
2.013
6
2.012
7
2.010
8
2.009
9
2.007
10
2.006
11
2.005
12
2.004
13
2.003
14
2.001
14.5
2.001
If the output voltage comes close to the expected value,
check the VID inputs at the device pins. The part is factory set to respond properly to the VID inputs.
Shut down at too low a current can be caused by an
inappropriate value of the sense resistor. See the “Sense
Resistor” section.
Poor load regulation can be due to many causes. Check
the voltages and signals at the critical pins.
The VREF pin should be at the voltage set by the VID
pins. If the power supply pins and the VID pins are correct the VREF should have the correct voltage.
Load Regulation 0.5-14.5A
0.90%
APPLICATION NOTE
AN53
The DC output voltage is measured from minimum to maximum load current. Ideally, the RC5051 should show about
20mV droop between these two limits. Note that the measurement must be taken directly on the leads of the output
capacitors, since there may otherwise be parasitic resistances
causing additional droop.
Output voltage load transients due to load
current step
Test
Condition Measured
Comment
Low to high
Current Step
(Vout =
2.00V)
0.5A-14.5A 96mV
Limit =
130mV
PASS
Figure 15
High to Low
Current Step
(Vout =
2.00V)
14.5A-0.5A 109mV
Limit =
130mV
PASS
Figure 16
Figure 16. High to Low Current Transiet Response
The output voltage transient response to the worst-case step
current load is measured using an Intel EMT tool. Again, the
measurement must be taken directly on the leads of the output capacitors, since there may otherwise be parasitic resistances and inductances causing additional droop.
Over-Current Limit
Using an electronic load, the output current should be slowly
increased until the converter hits over-current limit. Reaching this limit will be evident because the output voltage
quickly drops out of regulation as the current is increased.
Component Case Temperature
Figure 15. Low to High Current Transiet Response
Each of the major components should be measured for
excessive temperature, which can be indicative of a problem.
At the least, the RC5051, the MOSFETs and the schottky
diode should all be measured. The converter should be run at
its maximum load current until it reaches thermal stability, at
least 20 minutes, before the temperature measurements are
taken.
17
AN53
Notes
18
APPLICATION NOTE
APPLICATION NOTE
AN53
Notes
19
AN53
APPLICATION NOTE
LIFE SUPPORT POLICY
FAIRCHILD’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT DEVICES
OR SYSTEMS WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT OF FAIRCHILD SEMICONDUCTOR
CORPORATION. As used herein:
1. Life support devices or systems are devices or systems
which, (a) are intended for surgical implant into the body,
or (b) support or sustain life, and (c) whose failure to
perform when properly used in accordance with
instructions for use provided in the labeling, can be
reasonably expected to result in a significant injury of the
user.
2. A critical component in any component of a life support
device or system whose failure to perform can be
reasonably expected to cause the failure of the life support
device or system, or to affect its safety or effectiveness.
www.fairchildsemi.com
7/21/98 0.0m 001
Stock#AN30000053
Ó 1998 Fairchild Semiconductor Corporation