7 2004 Semtech Corp. www.semtech.com
POWER MANAGEMENT
SC4150
With the input bias current of the UV and OV comparators
in the range of 20-30nA, let’s choose the R1 to be
562kΩ. This yields the values of R2=9.31kΩ and R3 =
10.2kΩ. With these values the accuracy is about 1%
which is quite acceptable for those functions.
Resistor R4 sets the over-current trip. To choose R4,
the user must determine the level of the current where
it should trip. As a rule of thumb, the over-current is set
to be 200-300% of the nominal value. In our case, we
assumed this value to be 5A.
Considering the minimum trip voltage is 50mV the value
of R4 is 50mV ÷ 5A = 10 mΩ.
The tolerance of this resistor is usually price driven and
5% is an adequate range of accuracy.
The actual position and layout of the circuitry around the
sense resistor R4 is critical to avoid a false over-current
tripping. The trace routing between R4 and SC4150
should be as short as possible and wide enough to handle
the maximum current with zero current in the sense lines
– ideally “Kelvin” like.
Additionally, there is a short delay circuit at the
comparator to filter out unwanted noise and otherwise
induced transients.
Inrush Current
is being controlled by the R5C3 network
and swamping capacitor C2.
When a board is plugged into a live backplane, the input
bulk capacitance of the board’s power supply produces
large current transients due to the rush of the currents
charging those capacitors. The main feature of the
SC4150 is to provide an orderly and well-controlled inrush
current.
Since the minimum trip voltage is 50mV, let’s choose
the inrush current to be 3A.
Imax = Cload · ∆Vmax /dt
dt = Cload · ∆Vmax /Imax = 150µF · 70V / 3A = 3.5ms
This would be the minimum time for the gate voltage
plateau during which the Vdd linearly decreases
maintaining 3A charge current of the Cload.
The inrush can be calculated using the following equation:
IMAX = (50µA • CLOAD) / C3
With the values shown in the schematic the actual inruch
current will be about 2A, which is within the limits we
have chosen.
Resistor R5 will produce a time constant which prevents
Q1 from turning on when power is initially applied and
the circuit is not ready to actively pull the gate low. It’s
value is not critical and 18k ensures the adequate delay.
The value of C2 is chosen to prevent false turn-on of the
FET due to the current flowing via C3 into the gate of the
FET when the circuit initially connects to the power source.
Capacitors C2 and C3 form a divider from Vin to GND.
C2 must keep the initial voltage at the gate below Vth
minimum.
For the typical FET, this threshold is around 1V to 2V,
therefore C2 = 100 • C3 will keep gate voltage at 0.7V,
even at the ”worst” case of Vin = 70V.
The choice of the Q1 is quite straightforward and is guided
mostly by thermal considerations due to the power
dissipation in the steady state.
For instance, in our case, the nominal current is 2A, the
power dissipation due to the conducting losses will be
Pdis = Inom² • Rds_on.
The MOSFET should be able to withstand Vdss ≥ 100V
with continuous drain current Id ≥ 6A. Device SUD06N10
or similar fits this application. It has an Rds_on = 0.2Ω,
and will dissipate
Pdis = 2² • 0.2 = 0.8W, which can be handled by this
DPAK device.
If there is a consideration of reducing the temperature
of the MOSFET then the lower Rds_on device should be
chosen or a different style (D2PAK) which has lower
Junction-to-Ambient thermal characteristics.
The R6
has a function of dumping high frequency
oscillations. The value of it is not critical and can be in
the range of 5Ω to 20Ω.
Applications Information (Cont.)