versus power pulse duration and duty cycle. The
applications. Z
?/DIV>
is specified from junction-to-case for
power    MOSFETs    typically    used    in    telecom
applications.
The   following   example   provides   a   method   for
estimating the peak junction temperature of a power
MOSFET in determining if the MOSFET is suitable for
a
particular
application.
V
IN
(VDD VEE) = 48V, I
LIM
= 4.2A, t
FLT
is 20ms, and
the power MOSFET is the SUM110N10-09 (TO-263
package) from Vishay-Siliconix. This MOSFET has an
R
ON
of 9.5m& (T
J
= 25癈), the junction-to-case
thermal resistance (R
?J-C)
) is 0.4癈/W, junction-to-
ambient thermal resistance (R
?J-A)
) is 40癈/W, and the
Transient Thermal Impedance Curve is shown in
Figure 8. Consider, say, the MOSFET is switched on
at time t1 and the steady-state load current passing
through the MOSFET is 3A. At some point in time
after t1, at time t2, there is an unexpected short-circuit
applied to the load, causing the MIC2589/MIC2595
controller to adjust the GATE output voltage and
regulate the load current for 20ms at the programmed
current limit value, 4.2A in this example. During this
short-circuit load condition, the dissipation in the
MOSFET is calculated by:
P
D
(short) = V
DS
?I
LIM
; V
DS
= 0V (-48V) = 48V
P
D
(short) = 48V ?4.2A = 201.6W for 20ms.
At first glance, it would appear that a very hefty
MOSFET   is   required   to   withstand   this   extreme
overload condition. Upon further examination, the
calculation   to   approximate   the   peak   junction
temperature is not a difficult task. The first step is to
determine   the   maximum   steady-state   junction
temperature, then add the rise in temperature due to
the maximum power dissipated during a transient
overload caused by a short circuit condition. The
equation   to   estimate   the   maximum   steady-state
junction temperature is given by:
T
J
(steady-state) E T
C
(max) + T
J
(1)
T
C
(max) is the highest anticipated case temperature,
prior to an overcurrent condition, at which the
MOSFET will operate and is estimated from the
following equation based on the highest ambient
temperature of the system environment.
T
C
(max) = T
A
(max) + P
D
?(R
?J-A)
R
?J-C)
)
(2)
Lets assume a maximum ambient of 60癈. The
power dissipation of the MOSFET is determined by
the current through the MOSFET and the ON
resistance (I
2
R
ON
), which we will estimate at 17m&
(specification given at T
J
= 125癈). Using our
example information and substituting into Equation 2,
T
C
(max)     = 60癈+[((3A)
2
?7m&)?400.4)癈/W]
= 66.06癈
Substituting the variables into Equation 1, T
J
is
determined by:
T
J
(steady-state) ET
C
(max)+[R
O N
+(T
C
(max)T
C
)(0.005)
?(R
O N
)][I
2
?R
?(J -A )
R
?( J- C)
)]
E 66.06癈+[17m&+(66.06癈25癈)(0.005/癈)
?17m&)][(3A)
2
?400.4)癈/W]
E 66.06癈 + 7.30癈
E 73.36癈
Since this is not a closed-form equation, getting a
close approximation may take one or two iterations.
On the second iteration, start with T
J
equal to the
value calculated above. Doing so in this example
yields;
T
J
(steady-state) E66.06癈+[17m&+(73.36癈
-25癈)?0.005/癈)
?17m&)][(3A)
2
?400.4)]癈/W
E73.62癈
Another iteration shows that the result (73.63癈) is
converging quickly, so well estimate the maximum
T
J(steady-state)
at 74癈.
The use of the Transient Thermal Impedance Curves
is necessary to determine the increase in junction
temperature associated with a worst-case transient
condition. From our previous calculation of the
maximum power dissipated during a short circuit
event for the MIC2589/MIC2595, we calculate the
transient junction temperature increase as:
T
J
(transient) = P
D
(short) ?R
?J-C)
?Multiplier
(3)
Assume the MOSFET has been on for a long time
several minutes or more and delivering the steady-
state load current of 3A to the load when the load is
short circuited. The controller will regulate the GATE
output voltage to limit the current to the programmed
value of 4.2A for 20ms before immediately shutting off
the output. For this situation and almost all hot swap
applications, this can be considered a single pulse
event as there is no significant duty cycle. From
Figure 8, find the point on the X-axis (Square-Wave
Pulse Duration) for 25ms, allowing for a 25% margin