LT3748
21
3748fa
LPRI≤VIN(MIN)(VOUT+VF(DIODE))NPS/(fSW(MIN)ILIM
((VOUT + VF(DIODE)) NPS + VIN(MIN)))
LPRI ≥ (VOUT + VF(DIODE)) RSENSE 400ns NPS/15mV
LPRI ≥ VIN(MAX) RSENSE 200ns/15mV
For this application, the primary inductance with a 2:1
transformer and a 0.016Ω sense resistor for an 6.25A
current limit is bounded by the minimum desired switch-
ing frequency and the minimum off time requirement to
be between 9.6μH and 11.5μH. Looking at Table 1, there
are no transformers that fit that exact requirement. For the
sake of prototyping, a transformer with slightly less than the
desired primary inductance is selected with the PA3177NL.
The application will need to be tested thoroughly for sta-
bility at higher input voltages and when the current limit
is at a minimum (in the middle of the output load range).
The easiest solution to ease the requirement on minimum
on-time is to reduce the maximum VIN voltage although
alternatively NPS could be increased at the expense of ef-
ficiency (and requiring a more thorough redesign).
4. Select a MOSFET Switch
The selected 2:1 transformer requires a nominal 55V rating
on the MOSFET switch, assuming no leakage inductance.
However, even a small amount of leakage inductance may
cause the drain to ring to double the anticipated voltage,
and generally this needs to be verified in the final design.
However, at currents below 10A it is fairly easy to find a
MOSFET with sufficiently low RDS(ON) to be a very small
contributor to maximum load efficiency losses while
similarly having a low enough QG to require minimum
current and minimal losses when driving the MOSFET at
lighter loads. Also, while considering the efficiency gains
and losses with a given MOSFET, it is important to real-
ize that a trade-off in RDS(ON) for VDS(MAX) may backfire
if a snubber needs to be added to the circuit to meet the
voltage requirements and dissipates more energy than the
difference in switch resistance. For that reason, a Vishay
Si7738 is selected to give lots of margin with its 150V
rating. The RMS current in the MOSFET can be calculated,
squared and multiplied by the RDS(ON) to calculate losses
and the current required to drive the FET at frequency can
be determined, by the following equations:
IMOSFET(RMS) = √ILIM2 D/3
IINTVCC = fSW QG
PINTVCC = IINTVCC (VIN – VINTVCC)
In this application the MOSFET RMS current at maximum
load is about 2.7A, which into the 0.038Ω RDS(ON) will be
0.28W, or on the order of 2% loss in efficiency. Assuming
that the maximum operating frequency is around four
times higher than the maximum load frequency (at about a
quarter the output load) and reading the approximate QG at
7V operation from the Vishay data sheet, the approximate
INTVCC current is likely close to 8mA, dissipating 0.04W
when the load is on the order of 2.5W, or less than 2%,
and much less at maximum load.
5. Select the Output Diode
The output diode reverse voltage, as calculated earlier, is the
first important specification for the output diode. As with
the MOSFET, choosing a diode with enough margin should
preclude the use of a snubber. The second criterion is the
power requirement of the diode which is more difficult to
correctly ascertain—some manufacturers give direct data
about power dissipation versus duty cycle, which can be
used with the data from the table to determine. To avoid
using a snubber, a diode with a 60V reverse-bias capabil-
ity and minimal forward drop was selected—in this case,
the Diodes Inc. SBR 8U60P5. In this particular application
where maximizing efficiency is the goal, minimizing the
maximum voltage requirement on VIN may allow the use
of a diode with a lower reverse bias rating and a lower
forward drop which could further increase efficiency. Al-
ternatively, if no efficient diode is available for a particular
reverse bias rating, it may be more beneficial to increase
the windings ratio until a diode with low forward drop can
be selected and then reevaluate whether that solution with
higher RMS diode current is beneficial.
APPLICATIONS INFORMATION