Industry-Standard Pinouts
4.5 to 18V Operating Range
5A Peak Sink/Source at V
4.3A Sink / 2.8A Source at V
DD
= 12V
= 6V
OUT
Choice of TTL or CMOS Input Thresholds
Three Versions of Dual Independent Drivers:
- Dual Inverting + Enable (FAN3223)
- Dual Non-Inverting + Enable (FAN3224)
- Dual-Inputs (FAN3225)
Internal Resistors Turn Driver Off If No Inputs
MillerDrive™ Technology
12ns / 9ns Typical Rise/Fall Times with 2.2nF Load
Typical Propagation Delay Under 20ns Matched
within 1ns to the Other Channel
Double Current Capability by Paralleling Channels
8-Lead 3x3mm MLP or 8-Lead SOIC Package
Rated from –40°C to +125°C Ambient
Applications
Switch-Mode Power Supplies
High-Efficiency MOSFET Switching
Synchronous Rectifier Circuits
DC-to-DC Converters
Motor Control
Description
The FAN3223-25 family of dual 4A gate drivers is
designed to drive N-channel enhancement-mode
MOSFETs in low-side switching applications by
providing high peak current pulses during the short
switching intervals. The driver is available with either
TTL or CMOS input thresholds. Internal circuitry
provides an under-voltage lockout function by holding
the output LOW until the supply voltage is within the
operating range. In addition, the drivers feature matched
internal propagation delays between A and B channels
for applications requiring dual gate drives with critical
timing, such as synchronous rectifiers. This also
enables connecting two drivers in parallel to effectively
double the current capability driving a single MOSFET.
The FAN322X drivers incorporate MillerDrive™
architecture for the final output stage. This bipolarMOSFET combination provides high current during the
Miller plateau stage of the MOSFET turn-on / turn-off
process to minimize switching loss, while providing railto-rail voltage swing and reverse current capability.
The FAN3223 offers two inverting drivers and the
FAN3224 offers two non-inverting drivers. Each device
has dual independent enable pins that default to ON if
not connected. In the FAN3225, each channel has dual
inputs of opposite polarity, which allows configuration as
non-inverting or inverting with an optional enable
function using the second input. If one or both inputs are
left unconnected, internal resistors bias the inputs such
that the output is pulled LOW to hold the power
MOSFET OFF.
Related Resources
AN-6069 — Application Review and Comparative
Evaluation of Low-Side Gate Drivers
1. Estimates derived from thermal simulation; actual values depend on the application.
2. Theta_JL (
thermal pad) that are typically soldered to a PCB.
3. Theta_JT (
held at a uniform temperature by a top-side heatsink.
4. Theta_JA (Θ
The value given is for natural convection with no heatsink using a 2S2P board, as specified in JEDEC standards JESD51-2,
JESD51-5, and JESD51-7, as appropriate.
5. Psi_JB (
application circuit board reference point for the thermal environment defined in Note 4. For the MLP-8 package, the board
reference is defined as the PCB copper connected to the thermal pad and protruding from either end of the package. For the
SOIC-8 package, the board reference is defined as the PCB copper adjacent to pin 6.
6. Psi_JT (
the center of the top of the package for the thermal environment defined in Note 4.
): Thermal resistance between the semiconductor junction and the bottom surface of all the leads (including any
JL
): Thermal resistance between the semiconductor junction and the top surface of the package, assuming it is
JT
): Thermal resistance between junction and ambient, dependent on the PCB design, heat sinking, and airflow.
JA
): Thermal characterization parameter providing correlation between semiconductor junction temperature and an
JB
): Thermal characterization parameter providing correlation between the semiconductor junction temperature and
Stresses exceeding the absolute maximum ratings may damage the device. The device may not function or be
operable above the recommended operating conditions and stressing the parts to these levels is not recommended.
In addition, extended exposure to stresses above the recommended operating conditions may affect device reliability.
The absolute maximum ratings are stress ratings only.
Symbol Parameter Min. Max. Unit
VDD VDD to PGND -0.3 20.0 V
VEN ENA and ENB to GND GND - 0.3 VDD + 0.3V
VIN INA, INA+, INA–, INB, INB+ and INB– to GND GND - 0.3 VDD + 0.3V
V
OUTA and OUTB to GND GND - 0.3 VDD + 0.3V
OUT
TL Lead Soldering Temperature (10 Seconds) +260 ºC
TJ Junction Temperature -55 +150 ºC
T
Storage Temperature -65 +150 ºC
STG
Recommended Operating Conditions
The Recommended Operating Conditions table defines the conditions for actual device operation. Recommended
operating conditions are specified to ensure optimal performance to the datasheet specifications. Fairchild does not
recommend exceeding them or designing to Absolute Maximum Ratings.
Each member of the FAN322x driver family consists of
two identical channels that may be used independently
at rated current or connected in parallel to double the
individual current capacity. In the FAN3223 and
FAN3224, channels A and B can be enabled or disabled
independently using ENA or ENB, respectively. The EN
pin has TTL thresholds for parts with either CMOS or
TTL input thresholds. If ENA and ENB are not
connected, an internal pull-up resistor enables the driver
channels by default. ENA and ENB have TTL thresholds
in parts with either TTL or CMOS INx threshold. If the
channel A and channel B inputs and outputs are
connected in parallel to increase the driver current
capacity, ENA and ENB should be connected and
driven together.
The FAN322x family offers versions in either TTL or
CMOS input thresholds. In the FAN322xT, the input
thresholds meet industry-standard TTL-logic thresholds
independent of the V
voltage, and there is a
DD
hysteresis voltage of approximately 0.4V. These levels
permit the inputs to be driven from a range of input logic
signal levels for which a voltage over 2V is considered
logic HIGH. The driving signal for the TTL inputs should
have fast rising and falling edges with a slew rate of
6V/µs or faster, so a rise time from 0 to 3.3V should be
550ns or less. With reduced slew rate, circuit noise
could cause the driver input voltage to exceed the
hysteresis voltage and retrigger the driver input, causing
erratic operation.
In the FAN322xC, the logic input thresholds are
dependent on the V
logic rising edge threshold is approximately 55% of V
level and, with VDD of 12V, the
DD
DD
and the input falling edge threshold is approximately
38% of VDD. The CMOS input configuration offers a
hysteresis voltage of approximately 17% of V
. The
DD
CMOS inputs can be used with relatively slow edges
(approaching DC) if good decoupling and bypass
techniques are incorporated in the system design to
prevent noise from violating the input voltage hysteresis
window. This allows setting precise timing intervals by
fitting an R-C circuit between the controlling signal and
the IN pin of the driver. The slow rising edge at the IN
pin of the driver introduces a delay between the
controlling signal and the OUT pin of the driver.
Static Supply Current
In the IDD (static) typical performance characteristics
(Figure 12 - Figure 14 and Figure 19 - Figure 21) , the
curve is produced with all inputs/enables floating (OUT
is low) and indicates the lowest static I
tested configuration. For other states, additional current
flows through the 100k resistors on the inputs and
outputs shown in the block diagram of each part (see Figure 5 - Figure 7). In these cases, the actual static I
current is the value obtained from the curves plus this
additional current.
current for the
DD
DD
MillerDrive™ Gate Drive Technology
FAN322x gate drivers incorporate the MillerDrive™
architecture shown in Figure 47. For the output stage, a
combination of bipolar and MOS devices provide large
currents over a wide range of supply voltage and
temperature variations. The bipolar devices carry the
bulk of the current as OUT swings between 1/3 to 2/3
VDD and the MOS devices pull the output to the HIGH or
LOW rail.
The purpose of the MillerDrive™ architecture is to
speed up switching by providing high current during the
Miller plateau region when the gate-drain capacitance of
the MOSFET is being charged or discharged as part of
the turn-on / turn-off process.
For applications that have zero voltage switching during
the MOSFET turn-on or turn-off interval, the driver
supplies high peak current for fast switching even
though the Miller plateau is not present. This situation
often occurs in synchronous rectifier applications
because the body diode is generally conducting before
the MOSFET is switched ON.
The output pin slew rate is determined by V
voltage
DD
and the load on the output. It is not user adjustable, but
a series resistor can be added if a slower rise or fall time
at the MOSFET gate is needed.
V
DD
Input
stage
V
OUT
Figure 47. MillerDrive™ Output Architecture
Under-Voltage Lockout
The FAN322x start-up logic is optimized to drive
ground-referenced N-channel MOSFETs with an undervoltage lockout (UVLO) function to ensure that the IC
starts up in an orderly fashion. When V
below the 3.9V operational level, this circuit holds the
output LOW, regardless of the status of the input pins.
After the part is active, the supply voltage must drop
0.2V before the part shuts down. This hysteresis helps
prevent chatter when low V
supply voltages have
DD
noise from the power switching. This configuration is not
suitable for driving high-side P-channel MOSFETs
because the low output voltage of the driver would turn
the P-channel MOSFET ON with VDD below 3.9V.
To enable this IC to turn a device ON quickly, a local
high-frequency bypass capacitor C
ESL should be connected between the VDD and GND
pins with minimal trace length. This capacitor is in
addition to bulk electrolytic capacitance of 10µF to 47µF
commonly found on driver and controller bias circuits.
A typical criterion for choosing the value of C
keep the ripple voltage on the V
is often achieved with a value ≥20 times the equivalent
load capacitance C
, defined here as Q
EQV
Ceramic capacitors of 0.1µF to 1µF or larger are
common choices, as are dielectrics, such as X5R and
X7R with good temperature characteristics and high
pulse current capability.
If circuit noise affects normal operation, the value of
may be increased to 50-100 times the C
C
BYP
C
may be split into two capacitors. One should be a
BYP
larger value, based on equivalent load capacitance, and
the other a smaller value, such as 1-10nF mounted
closest to the VDD and GND pins to carry the higher
frequency components of the current pulses. The
bypass capacitor must provide the pulsed current from
both of the driver channels and, if the drivers are
switching simultaneously, the combined peak current
sourced from the C
would be twice as large as when
BYP
a single channel is switching.
with low ESR and
BYP
is to
supply to ≤5%. This
DD
BYP
GATE/VDD
EQV
.
, or
The FAN322x is compatible with many other
industry-standard drivers. In single input parts with
enable pins, there is an internal 100k resistor tied
to VDD to enable the driver by default; this should
be considered in the PCB layout.
The turn-on and turn-off current paths should be
minimized, as discussed in the following section.
Figure 48 shows the pulsed gate drive current path
when the gate driver is supplying gate charge to turn the
MOSFET ON. The current is supplied from the local
bypass capacitor, C
the MOSFET gate and to ground. To reach the high
peak currents possible, the resistance and inductance in
the path should be minimized. The localized C
to contain the high peak current pulses within this driverMOSFET circuit, preventing them from disturbing the
sensitive analog circuitry in the PWM controller.
The FAN3223-25 family of gate drivers incorporates
fast-reacting input circuits, short propagation delays,
and powerful output stages capable of delivering current
peaks over 4A to facilitate voltage transition times from
under 10ns to over 150ns. The following layout and
connection guidelines are strongly recommended:
Keep high-current output and power ground paths
separate logic and enable input signals and signal
ground paths. This is especially critical when
dealing with TTL-level logic thresholds at driver
inputs and enable pins.
Keep the driver as close to the load as possible to
minimize the length of high-current traces. This
reduces the series inductance to improve highspeed switching, while reducing the loop area that
can radiate EMI to the driver inputs and
surrounding circuitry.
If the inputs to a channel are not externally
connected, the internal 100k resistors indicated
on block diagrams command a low output. In noisy
environments, it may be necessary to tie inputs of
an unused channel to VDD or GND using short
traces to prevent noise from causing spurious
output switching.
Many high-speed power circuits can be susceptible
to noise injected from their own output or other
external sources, possibly causing output retriggering. These effects can be obvious if the
circuit is tested in breadboard or non-optimal circuit
layouts with long input, enable, or output leads.
For best results, make connections to all pins as
short and direct as possible.
Figure 48. Current Path for MOSFET Turn-on
Figure 49 shows the current path when the gate driver
turns the MOSFET OFF. Ideally, the driver shunts the
current directly to the source of the MOSFET in a small
circuit loop. For fast turn-off times, the resistance and
inductance in this path should be minimized.
The FAN3225 truth table indicates the operational states
using the dual-input configuration. In a non-inverting
driver configuration, the IN- pin should be a logic LOW
signal. If the IN- pin is connected to logic HIGH, a disable
function is realized, and the driver output remains LOW
regardless of the state of the IN+ pin.
IN+ IN- OUT
0 0 0
0 1 0
1 0 1
1 1 0
In the non-inverting driver configuration in Figure 50, the
IN- pin is tied to ground and the input signal (PWM) is
applied to IN+ pin. The IN- pin can be connected to logic
HIGH to disable the driver and the output remains LOW,
regardless of the state of the IN+ pin.
Operational Waveforms
At power-up, the driver output remains LOW until the
voltage reaches the turn-on threshold. The
V
DD
magnitude of the OUT pulses rises with V
steady-state V
is reached. The non-inverting
DD
operation illustrated in Figure 52 shows that the output
remains LOW until the UVLO threshold is reached, then
the output is in-phase with the input.
until
DD
Figure 50. Dual-Input Driver Enabled,
Non-Inverting Configuration
In the inverting driver application in Figure 51, the IN+
pin is tied HIGH. Pulling the IN+ pin to GND forces the
output LOW, regardless of the state of the IN- pin.
Figure 51. Dual-Input Driver Enabled,
Inverting Configuration
Figure 52. Non-Inverting Start-Up Waveforms
For the inverting configuration of Figure 51, startup
waveforms are shown in Figure 53. With IN+ tied to
VDD and the input signal applied to IN–, the OUT
pulses are inverted with respect to the input. At powerup, the inverted output remains LOW until the V
DD
voltage reaches the turn-on threshold, then it follows the
input with inverted phase.
Gate drivers used to switch MOSFETs and IGBTs at
high frequencies can dissipate significant amounts of
power. It is important to determine the driver power
dissipation and the resulting junction temperature in the
application to ensure that the part is operating within
acceptable temperature limits.
The total power dissipation in a gate driver is the sum of
two components, P
P
= P
TOTAL
GATE
GATE
+ P
and P
DYNAMIC
Gate Driving Loss: The most significant power loss
results from supplying gate current (charge per unit
time) to switch the load MOSFET ON and OFF at
the switching frequency. The power dissipation that
results from driving a MOSFET at a specified gatesource voltage, V
switching frequency, F
P
= QG • VGS • FSW • n (2)
GATE
, with gate charge, QG, at
GS
SW
n is the number of driver channels in use (1 or 2).
Dynamic Pre-drive / Shoot-through Current: A
power loss resulting from internal current
consumption under dynamic operating conditions,
including pin pull-up / pull-down resistors, can be
obtained using the “I
(No-Load) vs. Frequency”
DD
graphs in Typical Performance Characteristics to
determine the current I
under actual operating conditions:
P
DYNAMIC
= I
• VDD • n (3)
DYNAMIC
Once the power dissipated in the driver is determined,
the driver junction rise with respect to circuit board can
be evaluated using the following thermal equation,
assuming
was determined for a similar thermal
JB
design (heat sinking and air flow):
T
= P
J
TOTAL
•
+ TB (4)
JB
where:
T
= driver junction temperature
J
= (psi) thermal characterization parameter
JB
relating temperature rise to total power
dissipation
T
= board temperature in location as defined in
B
the Thermal Characteristics table.
:
DYNAMIC
(1)
, is determined by:
drawn from VDD
DYNAMIC
To give a numerical example, if the synchronous
rectifier switches in the forward converter of Figure 54
are FDMS8660S, the datasheet gives a total gate
charge of 60nC at V
= 7V, so two devices in parallel
GS
would have 120nC gate charge. At a switching
frequency of 300kHz, the total power dissipation is:
P
= 120nC • 7V • 300kHz • 2 = 0.5W (5)
GATE
P
P
= 1.5mA • 7V • 2 = 0.021W (6)
DYNAMIC
= 0.52W (7)
TOTAL
The SOIC-8 has a junction-to-board thermal
characterization parameter of
= 42°C/W. In a
JB
system application, the localized temperature around
the device is a function of the layout and construction of
the PCB along with airflow across the surfaces. To
ensure reliable operation, the maximum junction
temperature of the device must be prevented from
exceeding the maximum rating of 150°C; with 80%
derating, T
would be limited to 120°C. Rearranging
J
Equation 4 determines the board temperature required
to maintain the junction temperature below 120°C:
T
= TJ - P
B,MAX
T
= 120°C – 0.52W • 42°C/W = 98°C (9)
B,MAX
TOTAL
•
(8)
JB
For comparison, replace the SOIC-8 used in the
previous example with the 3x3mm MLP package with
= 2.8°C/W. The 3x3mm MLP package could operate
JB
at a PCB temperature of 118°C, while maintaining the
junction temperature below 120°C. This illustrates that
the physically smaller MLP package with thermal pad
offers a more conductive path to remove the heat from
the driver. Consider tradeoffs between reducing overall
circuit size with junction temperature reduction for
increased reliability.
Package drawings are provided as a service to customers considering Fairchild components. Drawings may change in any manner
without notice. Please note the revision and/or date on the drawing and contact a Fairchild Semiconductor representative to verify or
obtain the most recent revision. Package specifications do not expand the terms of Fairchild’s worldwide terms and conditions,
specifically the warranty therein, which covers Fairchild products.
Always visit Fairchild Semiconductor’s online packaging area for the most recent package drawings:
Package drawings are provided as a service to customers considering Fairchild components. Drawings may change in any manner
without notice. Please note the revision and/or date on the drawing and contact a Fairchild Semiconductor representative to verify or
obtain the most recent revision. Package specifications do not expand the terms of Fairchild’s worldwide terms and conditions,
specifically the warranty therein, which covers Fairchild products.
Always visit Fairchild Semiconductor’s online packaging area for the most recent package drawings:
http://www.fairchildsemi.com/packaging/
(1.04)
DETAIL A
SCALE: 2:1
C
0.10 C
x 45°
GAGE PLANE
0.36
SEATING PLANE
Figure 58. 8-Lead SOIC
0.19
OPTION A - BEVEL EDGE
OPTION B - NO BEVEL EDGE
NOTES: UNLESS OTHERWISE SPECIFIED
A) THIS PACKAGE CONFORMS TO JEDEC
MS-012, VARIATION AA, ISSUE C,
B) ALL DIMENSIONS ARE IN MILLIMETERS.
C) DIMENSIONS DO NOT INCLUDE MOLD
FLASH OR BURRS.
D) LANDPATTERN STANDARD: SOIC127P600X175-8M.
E) DRAWING FILENAME: M08AREV13