NCP1395A/B
APPLICATION INFORMATION
The NCP1395A/B includes all necessary features to help
28 mA hysteresis current for the lowest consumption
and the B version slightly increases this current to
83 mA in order to improve the noise immunity.
build a rugged and safe switch−mode power supply
featuring an extremely low standby power. The below
bullets detail the benefits brought by implementing the
NCP1395A/B controller:
• Wide Frequency Range: A high−speed Voltage
Control Oscillator allows an output frequency
excursion from 50 kHz up to 1.0 MHz on A and B
outputs.
• Adjustable Fault Timer Duration: When a fault is
detected on the slow fault input or when the FB path is
broken, a timer starts to charge an external capacitor.
If the fault is removed, the timer opens the charging
path and nothing happens. When the timer reaches its
selected duration (via a capacitor on pin 6), all pulses
are stopped. The controller now waits for the
• Adjustable Deadtime: Due to a single resistor wired
to ground, the user has the ability to include some
deadtime, helping to fight cross−conduction between
the upper and the lower transistor.
discharge via an external resistor of pin 6 capacitor to
issue a new clean startup sequence with soft−start.
• Cumulative Fault Events: In the NCP1395A/B, the
timer capacitor is not reset when the fault disappears.
It actually integrates the information and cumulates
the occurrences. A resistor placed in parallel with the
capacitor will offer a simple way to adjust the
• Adjustable Soft−Start: Every time the controller
starts to operate (power on), the switching frequency is
pushed to the programmed maximum value and slowly
moves down toward the minimum frequency, until the
feedback loop closes. The soft−start sequence is
activated in the following cases: a) normal startup
b) back to operation from an off state: during hiccup
faulty mode, brown−out or temperature shutdown
(TSD). In the NCP1395A, the soft−start is not
activated back to operation from the fast fault input,
unless the feedback pin voltage reaches 0.6 V. To the
opposite, in the B version, the soft−start is always
activated back from the fast fault input whatever the
feedback level is.
discharge rate and thus the auto−recovery retry rate.
• Fast and Slow Fault Detection: In some application,
subject to heavy load transients, it is interesting to
give a certain time to the fault circuit, before
activating the protection. On the other hand, some
critical faults cannot accept any delay before a
corrective action is taken. For this reason, the
NCP1395A/B includes a fast fault and a slow fault
input. Upon assertion, the fast fault immediately stops
all pulses and stays in the position as long as the
driving signal is high. When released low (the fault
has gone), the controller has several choices: in the
A version, pulses are back to a level imposed by the
feedback pin without soft−start, but in the B version,
pulses are back through a regular soft−start sequence.
• Adjustable Minimum and Maximum Frequency
Excursion: In resonant applications, it is important to
stay away from the resonating peak to keep operating
the converter in the right region. Due to a single
external resistor, the designer can program its lowest
frequency point, obtained in lack of feedback voltage
(during the startup sequence or in short−circuit
conditions). Internally trimmed capacitors offer a
"3% precision on the selection of the minimum
switching frequency. The adjustable upper stop being
less precise to "15%.
• Skip Cycle Possibility: The absence of soft−start on
the NCP1395A fast fault input offers an easy way to
implement skip cycle when power saving features are
necessary. A simple resistive connection from the
feedback pin to the fast fault input, and skip can be
implemented.
• Low Startup Current: When directly powered from
the high−voltage DC rail, the device only requires
300 mA to startup. In case of an auxiliary supply, the
B version offers a lower startup threshold to cope with
a 12 V dc rail.
• Onboard Transconductance Op Amp: A
transconductance amplifier is used to implement
various options, like monitoring the output current and
maintaining it constant.
• Broken Feedback Loop Detection: Upon startup or
any time during operation, if the FB signal is missing,
the timer starts to charge a capacitor. If the loop is
really broken, the FB level does not grow up before
the timer ends counting. The controller then stops all
pulses and waits that the timer pin voltage collapses to
1.0 V typically before a new attempt to restart, via the
soft−start. If the optocoupler is permanently broken, a
hiccup takes place.
• Brown−Out Detection: To avoid operation from a
low input voltage, it is interesting to prevent the
controller from switching if the high−voltage rail is
not within the right boundaries. Also, when teamed
with a PFC front−end circuitry, the brown−out
detection can ensure a clean startup sequence with
soft−start, ensuring that the PFC is stabilized before
energizing the resonant tank. The A version features a
http://onsemi.com
14