Datasheet
19Maxim Integrated
Integrated, 4-Channel, High-Brightness LED
Driver with High-Voltage DC-DC Controller
MAX16814
Feedback Compensation
During normal operation, the feedback control loop reg-
ulates the minimum OUT_ voltage to 1V when LED string
currents are enabled during PWM dimming. When LED
currents are off during PWM dimming, the control loop
turns off the converter and stores the steady-state condi-
tion in the form of capacitor voltages, mainly the output
filter capacitor voltage and compensation capacitor
voltage. For the MAX16814A_ _ and the MAX16814U_
_, when the PWM dimming pulses are less than or equal
to 5 switching clock cycles, the feedback loop regulates
the converter output voltage to 95% of OVP threshold.
The worst-case condition for the feedback loop is when
the LED driver is in normal mode regulating the minimum
OUT_ voltage to 1V. The switching converter small-signal
transfer function has a right-half plane (RHP) zero for
boost configuration if the inductor current is in continuous
conduction mode. The RHP zero adds a 20dB/decade
gain together with a 90N-phase lag, which is difficult to
compensate.
The worst-case RHP zero frequency (f
ZRHP
) is calcu-
lated as follows:
For boost configuration:
2
LED MAX
ZRHP
LED
V (1 D )
f
2 LI
−
=
π× ×
For SEPIC and coupled-inductor boost-buck configura-
tions:
2
LED MAX
ZRHP
LED MAX
V (1 D )
f
2 LI D
−
=
π× × ×
where f
ZRHP
is in hertz, V
LED
is in volts, L is the induc-
tance value of L1 in henries, and I
LED
is in amperes. A
simple way to avoid this zero is to roll off the loop gain
to 0dB at a frequency less than one fifth of the RHP zero
frequency with a -20dB/decade slope.
The switching converter small-signal transfer function
also has an output pole. The effective output impedance
together with the output filter capacitance determines the
output pole frequency f
P1
that is calculated as follows:
For boost configuration:
LED
P1
LED OUT
I
f
2V C
=
×π× ×
For SEPIC and coupled-inductor boost-buck configurations:
LED MAX
P1
LED OUT
ID
f
2V C
×
=
×π× ×
where f
P1
is in hertz, V
LED
is in volts, I
LED
is in amperes,
and C
OUT
is in farads.
Compensation components, R
COMP
and C
COMP
, per-
form two functions. C
COMP
introduces a low-frequency
pole that presents a -20dB/decade slope to the loop
gain. R
COMP
flattens the gain of the error amplifier for
frequencies above the zero formed by R
COMP
and
C
COMP
. For compensation, this zero is placed at the
output pole frequency f
P1
so that it provides a -20dB/
decade slope for frequencies above f
P1
to the combined
modulator and compensator response.
The value of R
COMP
needed to fix the total loop gain at
f
P1
so that the total loop gain crosses 0dB with -20dB/
decade slope at 1/5 the RHP zero frequency is calcu-
lated as follows:
For boost configuration:
ZRHP CS LED
COMP
P1 COMP LED MAX
f RI
R
5 f GM V (1 D )
××
=
× × × ×−
For SEPIC and coupled-inductor boost-buck configura-
tions:
ZRHP CS LED MAX
COMP
P1 COMP LED MAX
f RI D
R
5 f GM V (1 D )
×××
=
× × × ×−
where R
COMP
is the compensation resistor in ohms,
f
ZRHP
and f
P2
are in hertz, R
CS
is the switch current-
sense resistor in ohms, and GM
COMP
is the transcon-
ductance of the error amplifier (600FS).
The value of C
COMP
is calculated as follows:
COMP
COMP Z1
1
C
2R f
=
π× ×
where f
Z1
is the compensation zero placed at 1/5 of
the crossover frequency that is, in turn, set at 1/5 of the
f
ZRHP
.
If the output capacitors do not have low ESR, the ESR
zero frequency may fall within the 0dB crossover fre-
quency. An additional pole may be required to cancel
out this pole placed at the same frequency. This is usu-
ally implemented by connecting a capacitor in parallel