An-703 Application Note: Driving High Power Leds
An-703 Application Note: Driving High Power Leds
An-703 Application Note: Driving High Power Leds
APPLICATION NOTE
One Technology Way • P.O. Box 9106 • Norwood, MA 02062-9106 • Tel: 781/329-4700 • Fax: 781/461-3113 • www.analog.com
REV. 0
AN-703
By using the circuit shown in Figure 1, the output current CCH, CCL, and RC compose the loop compensation net-
equals work of the closed loop. They need to be designed
carefully to ensure system stability and control speed.
VISET
I LED = The control (duty cycle, d) to inductor current (iL) transfer
25RCS
function is a constant given by
If the current sensing signal in the design is lower than
50 mV, a bias circuit, shown in Figure 2, is needed to fully i L (S ) DVIN
=
activate the synchronous functionality. By using a voltage d 25RCSfSWL
divider composed of RB1 and RB2, a bias voltage, VBIAS, is
generated at the CS+ pin. Thus, the sensing signal needed where:
to activate the synchronous rectification function is VBIAS D is the duty cycle of the steady state.
lower than the 50 mV. If VBIAS equals 50 mV, the synchro- fSW is the switching frequency of the power stage.
nous function is effective all the time. With this bias, the
output current is This is valid below half of the switching frequency and
with the assumption that the output voltage is constant.
VISET − 25VBIAS
I LED = It is obvious that a simple integrator is adequate to com-
25RCS
pensate the loop with infinite dc gain. To set a bandwidth
The current limit setting accuracy is obviously a little of fC(C), the compensation component CCL needs to be
lower with this added offset term. Also, the biasing gMDVIN
circuitry should draw minimum current from the VREF C CL =
25RCSfSWLω C
pin, and filtering capacitor C4 should now be designed
to make a small enough time constant (recommend where:
0.3/f SW ) with RB2 instead of R CS . Thus, C4 is much
gM = 0.002 is the transconductance of the current loop
smaller than 200 nF.
error amplifier.
The output capacitor, COUT, is optional. It smoothes the
RC can be set to 0.
voltage and current across the LEDs.
CCH can be set to be open.
The input capacitor, CIN, is needed to absorb the input ripple
current. Normally, two 10 F/25 V ceramic capacitors are LEDs are not critical to the driving transition, so the sys-
enough. tem loop bandwidth can be low. It is recommended that
the bandwidth is designed to be 1/30th of the switching
The inductor, L, determines the ripple current. It is recom-
frequency for the buck configuration.
mended that the ripple current be about one-third of the
nominal output current to optimize both the system size MOSFETs can be chosen based on their dc voltage/
and the efficiency. The peak inductor current is therefore current ratings, switching speed (gate charges), and
thermal capability.
I LPEAK = 1.15I LEDM
Note that the BSTREG regulator needs to provide the
where: entire gate-driving energy to drive both the high-side
MOSFET and the high-side driver itself. Because the
ILPEAK is the peak current of the inductor.
maximum output current, IBSTREG , of the BSTREG regu-
ILEDM is the maximum average current in the LEDs. lator is 3 mA, the gate charge, QG , of the high-side
The inductor should be chosen based on two criteria. MOSFET and the switching frequency should satisfy
First, the inductor saturation current should be larger
I BSTREG > QGfSW
than the maximum peak current, ILPEAK, in the inductor.
Second, the inductor’s maximum dc current rating should If more gate-driving capability is needed, use an external
be larger than the load dc average current, ILED. NPN transistor, QBST, in the emitter follower configura-
The recommended bootstrap capacitance is 10 nF. A tion (see Figure 2) to deliver more current.
0.5 A Schottky diode is recommended as the bootstrap
diode. This diode must be placed on the PCB as close to
the BST pin and BSTREG pin as possible.
–2– REV. 0
AN-703
VIN 7V–20V
C1 R1
22�F 10�
1 VCC SW 24
C6
10nF CIN
2 SYS– DRVH 23 SW1
R2
100k�
3 SYS+ BST 22
D1
MBR052 L VOUT = 0 TO VIN
4 ISYS BSTREG 21
C5
1�F LED1
5 LIMIT DRVL 20 SW2 COUT
CT LED2
6 CT PGND 19
ADP3806
7 SYNC CS+ 18
C2 C4
200nF RCS
100nF
8 REG CS– 17
C3
200pF TO SET THE OUTPUT CURRENT
9 REF ISET 16
RCLB RCLT
10 SD BATSEL 15
R3
CCL 51k�
11 COMP BAT 14
RC
CCH
12 LC AGND 13
����������
�� ��
���� ���
� ��� �� ��
��
���� ���
� ���� ���� �� ���
��
�����
� ���� ��� ��
����
�������� � ���������������
��
� ���� ������ �� ������
��
��� ����
� ����� ���� �� ��� ����
������
�� ��������������� ����
� �� ���� ��
������� ���
� ���� ��� ��
�� ��
����� ���
����� ���
� ��� ��� ��
�� ����������
����� ��������������
� ��� ���� ��
���� ����
�� �� ������ ��
�� ���
��� ����
�� ���� ��� ��
�� ��������������������
���
�� �� ���� ��
Figure 2. Biasing the Current Sensing Input to Accommodate Low Input Signal
REV. 0 –3–
AN-703
BOOST CONFIGURATION The output capacitor, COUT, is optional. Without it, the
If the output voltage is higher than the input power current flowing into the LEDs is discontinuous and fluc-
source voltage to power a series of LEDs, step-up con- tuates at a duty of (1 – D), where D is the duty cycle of
version is used. the PWM control. With the output capacitor, the current
Figure 3 shows the boost configuration using the ADP3806 in the LEDs can be shaped smoother. Appropriate output
to drive LEDs. This scheme utilizes the main driving out- capacitance is needed to achieve the required driving
put DRVH to control the main switch SW1, with the diode current pattern.
switch SW2 used for rectification. The current sensing resistor, RCS , needs to be large
To avoid overvoltage output, the voltage loop can be enough to generate sufficient signal for the controller.
connected to set the maximum allowable output volt- The maximum differential input voltage between CS+
age, VOUTM . Because the internal reference is 2.5 V, the and CS– is 160 mV.
maximum output voltage is set by CCH, CCL, and RC need to be designed carefully to ensure
system stability and control speed. This design is different
ROVPT + ROVPB
VOUTM = 2.5 from that of the buck configuration.
ROVPB
With this configuration, the output is always connected
Choose appropriate voltage divider resistance to set the to the input power supply via SW2 and the inductor, even
overvoltage protection point, VOUTM. when the controller is in shutdown mode. This may pose
some issues for specific applications.
C1 R1
22�F 10�
1 VCC SW 24
CIN
2 SYS– DRVH 23
R2
100k� L
3 SYS+ BST 22
ADP3806
7 SYNC CS+ 18
C2 C4
200nF RCS
100nF
8 REG CS– 17
C3 TO SET THE
200pF OUTPUT CURRENT
9 REF ISET 16
–4– REV. 0
AN-703
SEPIC CONFIGURATION The output capacitor, COUT, is optional.
The SEPIC configuration provides an output that can be CCH, CCL, and RC need to be designed carefully to ensure
either higher or lower than the input voltage at a cost system stability and control speed. This design is different
of an additional inductor and capacitor. A SEPIC con- from that of the buck configuration.
figuration using the ADP3806 to drive LEDs is shown in
Figure 4. Unlike the boost configuration, the output is isolated to the
input source when the controller is in shutdown mode.
The output stage is very similar to the boost configura-
tion, except that the rectification switch, SW2, conducts
when the main switch, SW1, is on.
����������� ����
�� ��
���� ���
� ��� �� ��
���
� ���� ���� ��
��
����� ��
� ���� ��� ��
�� ��� ����
� ���� ������ ��
��
��� ����
� ����� ���� �� ��� �� ����
�� ����
� �� ���� ��
�������
� ���� ��� ��
�� ��
����� ���
�����
� ��� ��� ��
�� ����������
����� ��������������
� ��� ���� ��
REV. 0 –5–
AN-703
CUK CONFIGURATION Similarly, the output capacitor, COUT, is optional.
The CUK configuration provides an inverted output that CCH, CCL, and RC need to be designed carefully to ensure
can be either higher or lower in magnitude than the input system stability and control speed. This design is different
voltage. Figure 5 shows the CUK configuration using the from that of the buck configuration.
ADP3806 to drive LEDs.
The output is isolated to the input source when the
Similar to the buck configuration, the output current is controller is in shutdown mode.
continuous, with fewer pulsating components compared
to both the boost and the SEPIC configurations.
C1 R1
22�F 10�
1 VCC SW 24
CIN
2 SYS– DRVH 23
R2
100k� L1
3 SYS+ BST 22
CR L2 VOUT
4 ISYS BSTREG 21
C5
1�F LED1
5 LIMIT DRVL 20 SW1 SW2 COUT
CT LED2
6 CT PGND 19
RCS
ADP3806
7 SYNC CS+ 18
C2 C4
100nF 200nF
8 REG CS– 17
C3
200pF TO SET THE OUTPUT CURRENT
9 REF ISET 16
–6– REV. 0
AN-703
BRIGHTNESS CONTROL current rating and lower at reduced current. To over-
Brightness control can be implemented easily by control- come this problem, a PWM brightness control scheme
ling the voltage at Pin ISET. Referring to the data sheet, can be used.
the output current, IOUT, is set by VISET (voltage at the ISET By periodically shorting the ISET pin down to AGND,
pin) by VISET shows a pattern of PWM, as does the output cur-
VISET rent, IOUT. With an open-drain MOSFET connected to the
IOUT = ISET pin, the brightness of the output can be controlled
25RCS
by a digital signal, which can be generated by a micro-
There are two basic ways to adjust the brightness con- controller.
trol. The first is to control the output current by changing It takes time for the controller to react to the control
VISET at the ISET pin. This can be done simply by using a signal, so the frequency of this brightness control PWM
potentiometer running off the 6 V VREG rail or 2.5 V VREF signal should be much lower than the bandwidth of the
reference. control loop, which is much slower than the switching
To minimize the board space, layout inflexibility, and frequency. The brightness PWM signal can be in the
reliability concern, users may consider using a digital range of 50 Hz to 500 Hz.
potentiometer, AD5228, rather than its mechanical Figure 7 shows a typical high brightness efficiency
counterpart for the brightness control. AD5228 is a low curve for an LED vs. its driving current. IOPT is the opti-
cost, 32-step, manual control digital potentiometer that mal driving current point with the highest light emitting
features a built-in debouncer and zero-scale/midscale efficiency. The efficiency optimized brightness control
selectable preset. As shown in Figure 6, an AD5228 with approach is to drive the LEDs with a PWM current pat-
two external push buttons adds in the circuit flexibility tern. The current changes from zero to IOPT at a duty
with brightness control. DBRI. This leads to the maximum light-emitting output
This control is simple but may be inefficient because at the continuous IOPT driving. When more light output
the efficiency of LEDs tends to be higher at their full is needed, raise the driving current.
REV. 0 –7–
AN-703
����������
�� ��
���� ���
� ��� �� ��
��
���� ���
� ���� ���� �� ���
��
�����
���� ��� ��
AN04636–0–1/06(0)
�
��
������ � ���������������
� ���� ������ ��
��
��� ����
� ����� ���� �� ��� ����
�� ����
� �� ���� ��
������� ���
� ���� ��� ��
�� ��
����� ���
����� ���
� ��� ��� ��
��
�����
� ��� ���� ��
����
�� �� ������ �� ���
��
���
�� ���� ��� ��
�� ��������������������
�� �� ���� ��
���
���
���
� �� ��� ��
���
�����������������
����
������ ��
���������������
� �� ��� �� �����
� � � ��
� ��� � ��
����������������������
© 2006 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners.
–8– REV. 0