First, find out the duty cycles. Plug the numbers into the duty
cycle equation and we get D1 = 0.75, and D2 = 0.33. Next,
follow the decision tree in Figure 8 to find out the values of d1,
d2 and d3. In this case, d1 = 0.5, d2 = D2 + 0.5 - D1 = 0.08,
and d3 = D1 - 0.5 = 0.25. Iav = IOUT1·D1 + IOUT2·D2 = 1.995A.
Plug all the numbers into the input ripple RMS current equa-
tion and the result is Iirrm = 0.77A.
RDC is the winding resistance of the inductor. RDS is the ON
resistance of the MOSFET switch.
Example:
VIN = 5V, VOUT1 = 3.3V, IOUT1 = 2A, VOUT2 = 1.2V, IOUT2 = 1.5A,
RDS = 170mΩ, RDC = 30mΩ. (IOUT1 is the same as I1 in the
input ripple RMS current equation, IOUT2 is the same as I2).
20200246
FIGURE 8. Determining d1, d2 and d3
CATCH DIODE SELECTION
In boards that have internal ground planes, extending the top-
layer thermal pad outside the body of the package to form a
"dogbone" shape offers little performance improvement.
However, for two-layer boards, the dogbone shape on the top
layer will provide significant help.
The catch diode should be at least 2A rated. The most stress-
ful operation for the diode is usually when the output is shorted
under high line. Always pick a Schottky diode for its lower
forward drop and higher efficiency. The reverse voltage rating
of the diode should be at least 25% higher than the highest
input voltage. The diode junction temperature is a main con-
cern here. Always validate the diode's junction temperature
in the intended thermal environment to make sure its ther-
mally derated maximum current is not exceeded. There are a
few 2A, 30V surface mount Schottky diodes available in the
market. Notice that diodes have a negative temperature co-
efficient, so do not put two diodes in parallel to achieve a lower
temperature rise. Current will be hogged by one of the diodes
instead of shared by the two. Use a larger package for that
purpose.
Predicting on paper with reasonable accuracy the junction
temperature of the LM26400Y in a real-world application is
still an art. Major factors that contribute to the junction tem-
perature but not directly associated with the thermal perfor-
mance of the LM26400Y itself include air speed, air
temperature, nearby heating elements and arrangement of
PCB copper connected to the DAP of the LM26400Y. The
θ
JA value published in the datasheet is based on a standard
board design in a single heating element mode and measured
in a standard environment. The real application is usually
completely different from those conditions. So the actual θJA
will be significantly different from the datasheet number. The
best approach is still to assign as much copper area as al-
lowed to the DAP and prototype the design.
THERMAL CONSIDERATIONS
Due to the low thermal impedance from junction to the die-
attach pad (or DAP, exposed metal at the bottom of the
package), thermal performance heavily depends on PCB
copper arrangement. The minimum requirement is to have a
top-layer thermal pad that is exactly the same size as the
DAP. There should be at least nine 8-mil thermal vias in the
pad. The thermal vias should be connected to internal ground
plane(s) (if available) and to a ground plane on the bottom
layer that is as large as allowed.
When prototyping the design, it is necessary to know the
junction temperature of the LM26400Y to assess the thermal
margin. The best way to measure the LM26400Y's junction
temperature when the board is working in its usual mode is to
measure the package-top temperature using an infrared ther-
mal imaging camera. Look for the highest temperature read-
ing across the case-top. Add two degrees to the measure-
ment result and the number should be a pretty good estimate
of the junction temperature. Due to the high temperature gra-
17
www.national.com
NSC [ National Semiconductor ]
