Datasheet

Data Sheet ADP1874/ADP1875
Rev. A | Page 23 of 44
As soon as the forward current through the lower side MOSFET
decreases to a level where
10 mV = I
Q2
× R
ON(Q2)
the zero-cross comparator (or I
REV
comparator) emits a signal to
turn off the lower side MOSFE T. From this point, the slope of the
inductor current ramping down becomes steeper (see Figure 76)
as the body diode of the lower side MOSFET begins to conduct
current and continues conducting current until the remaining
energy stored in the inductor has been depleted.
HS AND LS
IN IDLE MODE
10mV = R
ON
× I
LOAD
ZERO-CROSS COMPARATOR
DETECTS 10mV OFFSET AND
TURNS OFF LS
SW
LS
0A
I
LOAD
t
ON
ANOTHER
t
ON
EDGE IS
TRIGGERED WHEN V
OUT
FALLS BELOW REGULATION
09347-074
Figure 76. 10 mV Offset to Ensure Prevention of Negative Inductor Current
The system remains in idle mode until the output voltage drops
below regulation. A PWM pulse is then produced, turning on the
upper side MOSFET to maintain system regulation. The ADP1875
does not have an internal clock, so it switches purely as a hysteretic
controller as described in this section.
TIMER OPERATION
The ADP1874/ADP1875 employ a constant on-time architecture,
which provides a variety of benefits, including improved load
and line transient response when compared with a constant
(fixed) frequency current-mode control loop of comparable
loop design. The constant on-time timer, or t
ON
timer, senses
the high-side input voltage (V
IN
) and the output voltage (V
OUT
)
using SW waveform information to produce an adjustable one-
shot PWM pulse. The pulse varies the on-time of the upper side
MOSFET in response to dynamic changes in input voltage,
output voltage, and load current conditions to maintain output
regulation. The timer generates an on-time (t
ON
) pulse that is
inversely proportional to V
IN
.
IN
OUT
ON
V
V
Kt ×=
where K is a constant that is trimmed using an RC timer product
for the 300 kHz, 600 kHz, and 1.0 MHz frequency options.
C
R (TRIMMED)
VREG
t
ON
V
IN
I
SW
INFORMATION
09347-075
Figure 77. Constant On-Time Time
The constant on-time (t
ON
) is not strictly constant because it
varies with V
IN
and V
OUT
. However, this variation occurs in such
a way as to keep the switching frequency virtually independent
of V
IN
and V
OUT
.
The t
ON
timer uses a feedforward technique, which when applied
to the constant on-time control loop makes it a pseudo-fixed
frequency to a first-order approximation. Second-order effects,
such as dc losses in the external power MOSFETs (see the
Efficiency Consideration section), cause some variation in
frequency vs. load current and line voltage. These effects are
shown in Figure 23 to Figure 34. The variations in frequency
are much reduced compared with the variations generated if
the feedforward technique is not used.
The feedforward technique establishes the following relationship:
K
f
SW
1
=
where f
SW
is the controller switching frequency (300 kHz,
600 kHz, and 1.0 MHz).
The t
ON
timer senses V
IN
and V
OUT
to minimize frequency variation
as previously explained. This provides pseudo-fixed frequency
as explained in the Pseudo-Fixed Frequency section. To allow
headroom for V
IN
and V
OUT
sensing, adhere to the following
equations:
VREG V
IN
/8 + 1.5
VREG V
OUT
/4
For typical applications where VREG is 5 V, these equations are not
relevant; however, care may be required for lower VREG/VIN
inputs.