Driving LEDs - Micropower Direct

Application Guide
Driving LEDs
The first commercial Light Emitting Diode (LED) was introduced in the 1960’s. From its early beginnings as a low intensity
red light, the LED has emerged as a highly versatile component,
critical to a wide variety of applications.
A Little History
released energy becomes electromagnetic radiation. The color and visibility
of the emission is dependent upon its wavelength. Color that is visible to
humans ranges from red (longest wavelength) to violet (shortest wavelength).
Figure 2 illustrates the color wavelength and the relative sensitivity of the
human eye.
Researchers at Texas Instruments discovered in 1961 that applying an electric
current to a gallium arsenide (GaAs) junction, caused an infrared radiation
emission. They applied for and received a patent for the infrared LED. At
General Electric Company, the first LED to produce light within the spectrum
visible to the human eye (about 655 nm) was produced in 1962. Early LEDs
were not very practical for most applications due to their low intensity, lack
of color variety and high expense. They found use primarily as indicators.
The first commercially viable LEDs were produced by the Monsanto Company
starting in 1968. These were fabricated using gallium arsenide phosphide
(GaAsP). The most significant early application was as segments in alphanumeric displays. Fairchild Optoelectronics produced the first very low cost
LEDs in the early 1970’s.
Continuing research led to the introduction of more colors (green and yellow)
and a wider usable wavelength. In the 1980’s, high brightness LEDs using gallium aluminium arsenide phosphide (GaAlAsP) were introduced. These devices
were bright enough to begin replacing incandescent bulbs in automotive and
traffic applications. By 1990, gallium aluminium indium phosphide (GaAlInP)
was being used to produce “super bright” LEDs.
In the 1990’s researchers in Japan developed a method of producing gallium
nitride (GaN) P-N junctions in a production environment. Using GaN, very
high intensity blue LEDs were introduced. By adding indium (InGaN), a high
intensity green LED was produced. Finally, lighting quality white LEDs were
introduced. More recently, the power output and light output efficiencies
have been significantly increased.
Today, low cost, high brightness LEDs are available in all colors. Surfacemount LEDs have been introduced and are available in single-color, bicolor,
and tricolor models. Research continues into the fabrication process, packaging options, and performance improvements, which in turn increases the
applications for which LEDs are a viable alternative.
Basic Theory
LEDs are complex, PN junction semiconductors. The typical structure (and
electronic symbol) is shown in Figure 1. When forward biased, current
flows from the anode (or
P Side) to the cathode
(or N Side). As the current passes through the
device, charge carriers
(electrons & holes) are
injected into the junction. A recombination of
the charge carriers occurs
when an electron meets a
hole. The electron drops
to a lower energy level,
releasing energy in the
form of a photon. This
effect is called injection
In semiconductors fabFigure 1: LED Construction & Symbol
ricated from materials
such as geranium or silicon (typically used in signal processing), this energy
is released as heat. When materials such as gallium arsenide are used, the
Figure 2: Color Wavelength
LEDs are highly directional, monochromatic devices. They are typically
fabricated from gallium based crystals. These crystals are then doped with
various inorganic materials (aluminium, arsenide, phosphide, indium, etc)
to produce emissions in a narrow frequency range. In this way the fabrication process is used to produce die that emit distinct colors. This output is
given as the peak wavelength in nanometers (λnm). Process variations will
typically yield chips with an output range of approximately ±5 to ±10 nm.
Table 1 gives the typical materials used to fabricate LEDs to achieve different
colors/wavelength ranges.
Much of the light genTable 1: LED Materials & Colors
erated by an LED is
reflected back into
Wavelength (nM)
the device. This is
a result of the high Infrared
GaAs, AIGaAs
refractive index misRed
610 - 760
match between the
semiconductor mateGaAsP,
590 - 610
rials used in fabrica- Orange
tion and the surroundGaAsP,
ing air. One way to
570 - 590
improve light extraction is to add a curGaP, AIGaInP, AIGaP,
490 - 570
rent reflecting layer,
GaN, InGaN
as shown in Figure
InGaN, GaN
450 - 490
1. Also called a current blocker or Bragg
400 - 450
reflector (mirror),
this layer redirects Ultraviolet
light back out of the
device, significantly
increasing the light output
t t off the
th LED.
LEDs are produced in batches, which can lead to performance variations
caused by differences in raw materials, handling, processing, etc from one
lot to another. To help minimize the effects of any
inconsistencies, manufacturers use a “binning”
system. LEDs are sorted into groups according
to brightness, color, forward voltage, etc.
Driving LEDs
MicroPower Direct
LED Terminology
Ambient Temperature: The temperature of the
area surrounding an LED light source.
Anode: The “positive” terminal connection to
an LED.
Balance Resistors: Resistors connected in series
with LED strings to help balance the current in
parallel connections. Typically very small values, they are sometimes referred to as ballast
Beam Angle: The angle between two lines on
either side of the optical axis at a point where
the luminous intensity is 50% of the center beam
intensity. Typically ranges from 8° to 160°. High
(or wide) beam angle LEDs (>70°) with their
broader spread of light are useful in illumination applications. Low (or narrow) beam angle
LEDs are typically used in indicator applications
where a higher luminous intensity is required
(for improved visibility). Sometimes called view
angle, viewing angle or beam spread.
Beam Lumens: The total lumens contained
within a light beam.
Beam Spread: See Beam Angle.
Binning: LEDs are sorted as part of the manufacturing process to help minimize operating
tolerances. Sort criteria includes intensity, color,
forward voltage, etc.
Brightness: See Luminance.
Bulb: Typically used in reference to a lamp.
An “LED bulb” is a finished lamp assembly that
contains LEDs.
Candela: (cd) The luminous intensity of a light
source in a given direction. At a wavelength of
555 nanometers (green), one candela will have
a radiant intensity of 1/683 watt per steradian.
Cathode: The “negative” terminal connection
to an LED.
Color Temperature: A measurement that indicates the hue of a specific type of light source.
Warm color temperatures tend to enhance red/
orange, adding a yellow tint to white. They are
typically used in homes, restaurants, etc. Cool
color temperatures enhance blue, adding a bluish tint to white. They are often used in offices,
hospitals, etc. Given in kelvins
Dominant Wavelength: (λd) The wavelength (or
color) of an LED as perceived by the human eye.
Visible LEDs are typically specified by their dominant wavelength or color. Sometimes referred to
as hue wavelength or hue sensation.
Eye Sensitivity: A curve depicting the sensitivity
of the human eye as a function of wavelength
Field Angle: Similar to “Beam Angle”, but given
at a point where the luminous intensity is 10% of
the center beam intensity.
Foot-candle: The illumination on a one square
foot surface set one foot from a one candla light
source. Equal to one lumen per square foot.
Forward Current: (IF) The current that flows
through the LED semiconductor junction when
it is forward biased.
Forward Voltage: (VF) The voltage drop across
the LED semiconductor junction when it is forward biased.
lluminance: A measure of the intensity of light
on a surface. Measured in foot-candles or lux, it
is inversely proportional to area.
IP Code: The International Protection Code
rates electronic enclosures as to the degree of
protection provided against the intrusion of solid
objects, dust, water, etc. Also called the Ingress
Protection Rating.
LED Construction
LEDs are produced using a wafer deposition process. The materials used depend upon the type of
LED desired. Metal contacts are added through a
photoresist/evaporation process. Finally, the wafer
is sawn into individual LED chips (typically about
0.25 mm square).
After processing, the chip is ready for use in either
a through hole lamp or an SMD lamp. A through
hole lamp (also called
a leaded or radial lamp)
is illustrated in Figure
3. The chip is mounted
on a lead frame. It is
placed in a cup (or well)
at a part of the lead
frame called the anvil
(because of its shape).
The anvil is part of the
cathode lead. The well
is coated in highly reflective material to help
direct emitted light back
out of the package.
The chip is attached to
the cathode lead using
conductive epoxy (in
some configurations the
die may have a wire
bond connection to the
cathode off the top of
the chip). Gold bonding
wire is used to connect
the die to the anode
post of the lead frame.
An SMD type LED is illustrated in Figure 4. Here, the
lead frame has been partially enclosed in epoxy,
typically by an injection molding operation. Again,
the chip is mounted on a lead frame. It is attached
to the cathode lead by conductive epoxy, and a
wire bond is used for the connection to the anode.
As is the case with the radial bulb, the chip sits in
a cavity formed by highly reflective material that
helps increase light output by redirecting emissions
out of the package. The cavity is filled with an
epoxy resin that protects
the chip and acts as a lens
for the light output.
SMD packaging is sometimes
used for higher power LEDs.
As is the case with any high
power device, care must
be taken to safely remove
excess heat generated by
the power dissipated within
the chip. Otherwise, the
bulb may be damaged from
overheating. For high power
products, the die is typically
connected (thermally) to
a heat sink placed in the
bottom of the package. In
the actual application, this
integrated heat sink may
be connected to external
heat sinking or air flow. This
connection can be made via
plating on the PC board,
heat pipes (or vias), etc.
Figure 3: Radial LED Bulb
Surface Mount LEDs are also
cost-efficient solutions for
low-power, compact designs. The products come in
a variety of available color,
lens, and package types and
are highly durable.
The whole assembly is
encapsulated within an
epoxy lens. The encapsulant protects the chip
and wire bonds from
damage due to vibration
SMD LEDs are smaller than
or shock. The diffusion
leaded components. Beof the encapsulation
cause they are low profile
(set by adding glass parand mounted directly on
ticles to the epoxy) is
the PCB, they are somealso a factor in setting
times used with light guides
Figure 4: SMD LED Bulb
the viewing (or beam)
angle of the light generated by the chip. Other fac- (or pipes) to direct the light output as required by
tors are the shape & size of the chip; the shape & the application. Most manufacturers provide them
size of the reflector cup; and the distance between on tape & reel for use with high speed, automated
the chip and the top of the lens (set by extending assembly equipment.
the lead frame into the assembly).
Through hole LED lamps have been available for
many years. They are produced in a variety of
standard package sizes (typically 3 to 10 mm),
and colors. They are very reliable, offer robust
performance and low power consumption.
Through hole LED lamps are used in a wide variety
of applications. They are often found in outdoor
LED panels, front panel indicators, instrumentation
indicators and small area back lighting.
SMD LEDs are very small (making them an ideal
choice for space-limited applications); highly
resistant to shock and vibration and are very light
weight. They are used in a variety of applications, including automotive lighting, push button
backlighting, and front panel indicators. Their
small size & weight, combined with low power
consumption make them a good choice for mobile
equipment and the ability to package them on reels
makes them attractive to high volume applications
where reduced assembly cost is required.
Table 2: Comparison Of Lighting Technologies
Life Span
>60 kHrs
1.2 kHrs
Compact Fluorescent
8.0 kHrs
1.0 kHrs
Turn On
Semiconductor < 1 Sec Directional 2,600 - 10,000 k
< 1 Sec
2,300 - 3,300 k
Mercury Vapor < 60 Sec
4,000 - 8,000 k
2,500 - 3,000 k
<150 Sec
Powering LEDs
LEDs are current controlled semiconductor devices.
The intensity of the light output is proportional to
the current flowing
through the junction. Care must be
taken to observe
the correct polarity of the LED and
not to exceed the
maximum current
rating. In either
case, catastrophic
damage to the LED
may occur. The
voltage flowing in
the circuit has to be
enough to provide
the forward drop
Figure 5: Constant Voltage required by the LED
(or LED’s) connected.
Many existing LED applications use “off the shelf”
commercial power supplies for a power source,
typically due to cost and availability. These power
sources are overwhelmingly constant voltage
devices. A simple connection is shown in Figure
5. The power source for this circuit is the D102, a
standard (5V input, 9V output @ 300 mA) DC/DC
converter. To protect the LED, the resistor R1 is
used to limit current flow. The following equation
is used to calculate the correct value of R1:
R1 = VS - VF
Taking specifications from a typical LED datasheet,
we can use a forward voltage drop of 2.0V and an
optimum forward current is 300 mA. This would
yield a value for R1 of:
R1 = 9.0V - 2.0V = 233X
With a simple current limiting resister, the LED
could be powered and protected. However, the
limitations of this approach are soon apparent.
The voltage controlled power source is the main
problem. Any changes in the output load will
result in variations in the output current (as the
power source regulates the output voltage). Load
variations can be caused by any number of circuit
components. These can range from changes in the
power supply input, circuit changes over temperature, or variations in the LED forward voltage drop
(for multiple LED connections). Any changes in the
output current will cause a change in LED brightness. Significant changes could cause damage.
LEDs typically have a positive temperature coefficient (PTC). When in use, as the LED warms up
the forward voltage will start to drop. This will
cause the LED to draw more current which, in turn,
will increase its temperature. Uncontrolled, this
can lead to thermal runaway and a catastrophic
failure of the LED. Even short periods of operation
under conditions exceeding recommended operating temperature limits can significantly reduce the
operating lifetime of an LED.
In more complex LED circuits, such as parallel output connections, unbalanced voltages may cause
a variety of issues. These could include variations
in brightness or color shifting from one LED string
to another,unacceptable in most applications if
they are visible to the human eye. In the event
of a catastrophic failure of one LED, the whole
system could fail in a chain reaction.
A voltage controlled source can be effective for
low current applications where the input range to
the supply is tightly controlled. For more complex
applications or higher current requirements, it is
better to use a constant current source.
Constant Current Drivers
As the name implies, a constant current driver
regulates the output current level. For any changes
in input line, temperature or output load, the output current is maintained within a regulation band.
The output voltage level is varied to achieve this.
For designers of LED lighting systems, this often
meant a choice between expensive constant current output power supplies; or the use of linear
or switching regulators that typically used an
external sensing resistor to monitor and control
the current output.
Recently, many low cost, constant current drivers
have come on the market; including a full line offered by MicroPower Direct. Now available over
a wide range of power and with DC or AC inputs,
these units offer designers a quick, compact,
and economical solution to driving LED lights in
a variety of configurations for a wide range of
Figure 6 shows a
simplified connection of a constant
current driver. As
with most of these
devices, this driver
allows the user
to set the output
current to the desired level for the
specific application. In this case,
30 mA. So, for our
example, I OUT is Figure 6: Constant Current
equal to:
IOUT = IF = 30 mA
Once set, the driver will maintain the output current level to within a tight regulation band
Using An LED Driver
To illustrate LED driver connections, we will use
the LD24-08-300. This unit is a low cost DC/DC
driver with a constant current output. It is packaged in a small, encapsulated 0.8 x 0.4 case. The
specifications of this model are summarized below.
Figure 7: LD24-08-300 Specification Summary
Input Voltage Range
Max Input Voltage
Output Voltage Range
Output Current
Output Power
Analog Dimming
Adjust Voltage Range
Output Current Adjustment
Digital Dimming
Max Operation Frequency
Switch On Time
Switch Off Time
7 - 30
2 - 28
0.3 to 1.25
25 to 100
LED Terminology
Light Emitting Diode: (LED) A diode that emits
photons (as light) when forward biased.
LED Strip: LEDs that are attached to a flexible
PC board up to about 16 feet long and put on
reels. The user can than trim the strip to the
size required.
Lumen: A lumen is the luminous flux of light
produced by an LED that emits one candela
of luminous intensity over a solid angle of one
steradian (sr).
Lumen Maintenance: The ability of an LED light
to maintain intensity over time. A high power
LED will typically retain 70% of its intensity for
up to 50k hours.
Luminance: The luminous flux emitted or reflected from a source; in this case an LED. Given
in lumens (lm).
Luminosity Function: Established by the CIE,
this function approximates the average visual
sensitivity of the human eye to light of different
wavelengths. Two functions are defined. The
photopic luminosity function is used for everyday
light levels; while the scotopic luminosity function is used for poor light levels. Also called the
luminous efficiency function.
Luminous Efficacy: A measure of the effectiveness of a light source in converting electrical
energy into light. It’s the ratio of luminous flux to
power & is expressed as lumens per watt (lm/W).
Luminous Flux: (F) A measure of the total perceived power of a light source in all directions.
The measurement factors in the sensitivity of
the human eye by incorporating the luminosity
function. Expressed in lumens. Sometimes called
luminous power. See Luminosity Function.
Luminous Intensity: The perceived power emitted by a light source in a single direction. It is
the luminous flux per unit solid angle steradian
(sr). Expressed in candelas (cd).
Lux: (lx) The measure of light intensity, as
perceived by the human eye. One lux equals
one lumen per square meter.
Nanometer: (nm) A unit of length in the metric
system, equal to one billionth of a meter. Used
as a measure of the wavelength of light.
Operating Life: The number of hours an LED is
expected to be operational. For illumination
applications where light output is considered
critical, output degradation to 70% lumens is
typically used. For applications where light
output is not as critical (such as decorative
lighting), 50% is typically used. Given in hours.
Peak Wavelength: (λp) The single wavelength
where the radiometric emission spectrum of an
LED reaches its maximum
PWM: Pulse Width Modulation. A circuit that
varies the brightness of an LED by changing
the duty cycle of the output current of the
LED driver.
Radiant Flux: The total power of electromagnetic radiation (including infrared, ultraviolet,
and visible light) emitted from an LED. Measured in watts, it is also called radiant power.
Radiant Intensity: A measure of the intensity of electromagnetic radiation, defined as
power per unit solid angle. Given in watts per
Reverse Breakdown Voltage: Amount of
reverse bias that will cause a P-N junction to
break down and conduct in the reverse direction.
Web: www.micropowerdirect.com • Email: [email protected] • Tel: (781) 344 - 8226 • Fax: (781) 344 - 8481
Page 3
MicroPower Direct
Driving LEDs
Figure 8: Simplified Driver Connection
A simplified parallel connection using the LD2408-300 is shown in Figure 8 above.
The LD24-08-300 is a buck converter, the most
common type of DC/DC driver now available. With
a buck converter, the output voltage is always
slightly lower than the input. In this case, about 2V.
With an upper range limit set at 30V, it is capable
of driving LED strings with a combined forward
voltage drop of 28V maximum.
The other components shown are optional, to be
used dependent upon the requirements of the
specific application.
DC Input
The transient voltage suppressor (TVS) T1 is used
to meet the surge requirements of EN 61000-4-5.
The clamping voltage of the TVS must be <40
VDC maximum. This will prevent any surge from
exceeding the maximum input of the driver (40
VDC). Exceeding the maximum input rating could
damage the driver.
The Pi filter shown (C1, C2 and L1) will help to
meet conducted emission requirements. With the
addition of the filter, the unit should meet the
levels of EN 55015.
Figure 9: AC/DC Input Stage
AC Input
Our example of Figure 8 shows a DC/DC driver
connected to a stable DC input. Many applications,
for a variety of reasons, will require an AC input.
There are drivers available that integrate an AC
input into the package with the driver. Another
option is too add an AC/DC power supply to the
input stage of the circuit (as shown in Figure 9).
This is a distributed (or two-stage) connection.
circuit for setting the output current of the driver.
The output is varied by changing the voltage level
at the VADJ input (pin 2). Per the specifications,
the output current can be varied from 75 mA to
300 mA by changing the VADJ level from 0.30 VDC
to 1.25 VDC. Care must be taken not to exceed
1.25 VDC to avoid possible damage to the driver.
If the pin is left open, the output current is 300
mA (full), if it’s grounded, the driver shuts down.
In this connection, we take the AC line in (90 to 264
VAC) and connect it to the MPM-04S-12. This is a
miniature, 4W AC/DC power supply that provides
a tightly regulated 12 VDC output at 333 mA. The
12 VDC output powers the LED driver.
Our circuit has a 12 VDC input to the LED driver. To
insure that we do not exceed the 1.25 VDC limit
on the VADJ pin, a shunt regulator is connected in
parallel with the resistor network R2 and R3. Over
input levels of 5V to 30V, the shunt regulator (SR1)
will maintain the voltage across R2 and R3 at 2.5
VDC. By adjusting R3, the voltage level on the
VADJ pin is varied. The output current is equal to:
The two stage approach can simplify the safety
approval process (most AC/DC power supplies on
the market are approved to EN 60950) and may
increase design flexibility. Besides the output
power, other specifications to consider when
selecting the input AC/DC supply would include
input range, safety approvals, PFC rating (which
may be needed for various system energy ratings)
and operating temperature range.
Output Current Adjustment
The Figure 8 illustration shows a simple analog
R3 ) V
R2 + R3
0.08925 X (
IOut =
The VCNT used in the formula is the voltage level
used to set-up VADJ. In this instance, it is the 2.5
VDC level set by the regulator SR1. Quite often it
is VIN or some regulated bus level that is available.
For our example, we need an output current level
of 90 mA (30 mA for each of three output stacks).
The VADJ setting is equal to:
VADJ = IOUT X 0.372
To set the output at 90 mA, this gives us a VADJ
setting of:
VADJ = 0.09 X 0.372 = 0.375 VDC
We can also derive the VADJ level from the formula:
R3 X V
R2 + R3
We need to know what value to set R3 at to get a
0.375 VDC at the VADJ input. Since we know the
VADJ level required, we can now calculate a value
for R3 using the following formula:
R3 = R 2 X VADJ
Thus, the correct value of R3 is:
R3 =
10, 000 x 0.375
= 1.76 kX
2.5 - 0.375
So for our example, adjusting R3 to 1.76 kΩ will
set the VADJ level at 0.375V which will in turn set
the output current at 90 mA.
Due to component tolerances, rounding and a
slight non-linearity in the IOUT/VADJ curve of the
LD24-08-300, these formulas may not yield exact
results. However, with a little tweaking the results
should be satisfactory.
Figure 10 shows a slightly different approach using
two low cost, switching regulators. Working from
inputs that can range from 15 VDC to 32 VDC,the
top regulator (SR1) keeps the input to the LED
driver at 12 VDC.
The other regulator (SR2), driven off the same
input line maintains the control voltage at 5 VDC.
The resister network of R1 and R2 can now be used
to set the output current level of the LED driver.
The same equations we have just discussed are
still applicable to this circuit with the change of
VCNT from 2.5 VDC to 5 VDC.
Dimming LEDs
The circuits just discussed could also be used to
dim the LEDs. This is accomplished by simply lowering the driver output current below the specified
drive current for the LEDs being used. While this
method is common, it does not give the best results
for many applications.
An LED operates at its maximum efficiency when
operated at the rated drive current specified by
the manufacturer. Operating an LED at lower than
its rated forward current will not only decrease
the system efficiency, but may cause color (or
wavelength) shifting. In illumination applications,
this could cause visible changes to lighting.
A preferred method is using pulse width modulation (PWM). Since LEDs reach full light output
almost instantaneously, it is possible to change
the intensity level by rapidly turning the LED on/
off. By changing the duty cycle of the on/off time,
the perceived intensity of the light is varied up
or down. Keeping the frequency rate at greater
than 100 Hz will avoid any flicker that is visible to
the human eye.
Figure 10: Input Using Switching Regulators
Figure 11: Digital Dimming (PWM Dimming)
Figure 11 shows a simple method of achieving
digital (or PWM) dimming. Here, instead of using
a DC voltage level to set the output current level,
we are using a 555 timer to apply a series of pulses
to the VADJ input.
The 555 operates over a supply voltage range of
4.5 VDC to 15VDC. Here we have it connected to
the 12 VDC output of the switching regulator (this
is also the VIN of the LED driver). Care should be
taken to minimize ripple at the VCC input. Excess
ripple could cause timing errors.
The timer is connected for astable (free run) operation. The frequency is set by R1, R2 and C4. The
timing capacitor (C4) charges through R1 and D2.
When it reaches the level of 2/3 VCC, the discharge
pin (pin 7) goes low and C4 will discharge through
D1 and R2 to the internal discharge transistor.
When the C4 voltage drops to 1/3 VCC, the discharge
pin goes high and C4 begins to charge again. The
frequency is derived from the following formulas.
TON (τ) is equal to:
TON = 0.67 X R1 X C 4
TOFF is equal to:
TOFF = 0.67 X R 2 X C 4
The total period (Τ) is equal to:
Which gives us a frequency (f) of:
f= 1
And finally a duty cycle (D expressed as a decimal) of:
S R1 + R 2
For these examples we are ignoring the 0.6V drop
across the diodes. The diodes (D1 and D2) allow
duty cycles below 50% to be set. Diode D1 bypasses
R2 while C4 is charging. Diode D2 is optional (but
recommended), essentially blocking R2 during the
charge period. For our example, we want a 50%
duty cycle. To achieve this, we need to calculate
the correct value of R2. We can use the following
R 2 = R1 - R1
For our 50% duty cycle, this gives us:
R2 =
10, 000
- 10, 000 = 10 kX
Theoretically, this circuit will allow for duty
cycles over a range of approximately 5% to 95%.
If manual adjustment is desired, a potentiometer
may be substituted for R2 (with some adjustment
of the circuit).
The 555 timer is very accurate, so inaccuracies in
using these formulas are probably due to tolerances in the external components used. The timing capacitor (C4) should be a tantalum, mylar, or
equivalent (ceramic disc capacitors should not be
used). The size of C4 is generally not critical, but
it should be as low leakage as possible.
The timing resistors (R1 & R2) should be metal film.
In order to avoid excessive current flow through
Web: www.micropowerdirect.com • Email: [email protected] • Tel: (781) 344 - 8226 • Fax: (781) 344 - 8481
Page 5
MicroPower Direct
Driving LEDs
Series Connection Advantages:
• Circuit complexity is very low
• Each LED sees the same current
• High circuit efficiency (no need for balancing
the internal discharge transistor, it is recommended that R1 be at least 5 k⍀.
The timer also requires a minimum value of
current to operate the internal threshold
comparator (typically about 0.25 µA). Care
must be taken not to install values of R1 & R2
that would limit the threshold current to a level
that is insufficient to trip the comparator. To
calculate the maximum value of resistance,
use the following formula:
Series Connection Disadvantages:
• The required driver output voltage can
become very high for large LED strings
• Entire string fails if one LED opens
A shorted LED has little effect on the circuit
operation (light output will dim by 1/n, where n
equals the number of LEDs in the string). However,
if the LED string fails open for any reason (LED
failure, mechanical connection, etc), the whole
circuit fails.
The voltage at the threshold pin (pin 6) is 1/3
VCC, which in turn gives us a VCAP voltage of
2/3 VCC. Using this, we can calculate RMAX as
If timer accuracy over temperature is critical
to the application, external components with
a slight positive temperature coefficient should
be used. This will counteract the typically small
negative temperature coefficient of the timer and
cancel any timing drift over temperature.
The output of the timer is a high power totem pole
circuit. The bypass capacitor (C5) eliminates any
current spikes this causes on the VCC input. The
value of C5 is not critical and is typically between
0.01 µF and 10 µF. Howevr, C5 should be mounted
as close to the timer as possible.
The peak output voltage level of the timer is
equal to:
VPK = VCC - 1.7 VDC
For our circuit this equals:
VPK = 12 VDC - 1.7 VDC = 10.3 VDC
This gives us an output that is a series of pulses
with a 10.3 VDC peak value. The turn on/off rate
is 50%. To safely apply this signal to our driver/
LED circuit, the peak value must be reduced to
the correct level. In our example, this has already
been calculated at 0.375 VDC.
To achieve this, we are using a simple divider
network (R3 & R4).Using some earlier equations,
we can calculate the correct values for this diver
network. For VADJ, we have:
R4 X V
R 4 + R3
Once again, we need to know what value to set R4
at to get a 0.375 VDC at the VADJ input. Since we
know the VADJ level required, we can calculate a
value for R4 using the following formula:
R 4 = R3 X VADJ
Thus, the correct value of R4 is:
R4 =
Parallel Connection
Figure 13 illustrates a parallel (actually a seriesparallel) connection. The driver is now powering
fifteen LED’s connected in three parallel strings.
For this circuit, the driver IOUT is equal to:
12 - 8
= 16 MX
0.25 X 10 -6
30, 000 x 0.375
= 1.13 kX
10.3 - 0.375
Now, in our example, a pulse wave is applied to
the VADJ input of the driver. This signal has a 0.375
VDC peak amplitude and a 50% duty cycle. With
this control signal applied to the VADJ input, the
output of the driver will be a pulse train with a
50% duty cycle and a peak current of 30 mA. The
LEDs will appear to be running at 50% intensity
with no variation in color.
Connecting the LEDs to the driver is typically done
in series strings, or parallel strings. A series connection (as shown in Figure 12), is the simplest
and most common type.
Figure 12: Simplified Series Connection
Output Circuits
Series Connection
Driving multiple LEDs in series avoids uneven light
levels resulting from current variations. All of the
LEDs in the string see the same forward current,
insuring maximum brightness matching.
The output voltage of the driver is equal to:
Where VF is the specified LED forward voltage
and n is the number of LEDs in the string. In this
case, we have a typical VF of 2 VDC and five LEDs,
resulting in a 10 VDC output. Most DC/DC LED
drivers are step-down buck regulator types. Care
must be taken that the specified input range of
the driver exceeds the required output level by
an appropriate margin.
The output current of the driver is equal to:
All LEDs in the string see the same current. In our
example, this would be 30 mA.
Figure 13: Simplified Parallel Connection
IOUT = IF1 + IF2 + IF3
For our circuit, this equals:
IOUT = 30 mA + 30 mA + 30 mA = 90 mA
The driver output voltage required is still equal
to the total of the VF drops in one string (assuming the strings are balanced). In our case, this is
again 10 VDC.
The major advantage of using parallel strings is
the number of LEDs that can be powered without
exceeding the upper voltage limitation of the
driver. In this case, the LD24-08-300 has an upper
output voltage limitation of 28 VDC. The highest
number of LEDs with a 2V forward voltage drop
it could power in a serial connection would be
fourteen (with no guard band). Using a parallel
connection, the output voltage required is that of
one string (in this case about 10V), allowing our
driver to power many more LEDs.
The major problem with the parallel connection
is that small differences in circuit tolerances can
cause significant differences in the current drawn
by each string (or stack). This can lead to problems
ranging from differences in the perceived intensity
or color of the LEDs to catastrophic failure of one
or more stings.
Figure 14: Simplified Matrix Connection
The balancing resistors (RB1, RB2 & RB3) are used
to help compensate for current variations caused
by differences in the typical VF of the LED strings.
Small imbalances in the typical VF of the strings
could cause significant variance in the string current. The typical resistance value is small (<20⍀).
Our example circuit illustrated in Figure 8 (page
4) uses a current mirror to regulate the current
through the individual strings. This is a current sink
mirror, in which Q1 is connected as a diode and
controls the current flowing through Q2 and Q3. If
the transistors are well matched (for specifications
such as VBE and operation over temperature), the
current through each stack should be reasonably
close. To help maintain accuracy over temperature, the transistors need to be thermally connected. Mounting them to the same heat sink is a
common method of achieving this. The balancing
resistors (RB1, RB2 & RB3) are still used (partly to
compensate for small changes in VBE).
Parallel Connection Advantages:
• Ability to drive higher numbers of LEDs
Parallel Connection Disadvantages:
• Lower Efficiency
• Increased Circuit Complexity
• Low reliability
Low reliability (as configured in Figure 13) is caused
by increased risk of failure due to potential current variations. A shorted LED will cause increased
current to flow through the remaining LEDs in the
same string as the faulty LED. Since the total current is fixed by the driver output, this increase in
the faulty string will cause the other strings to dim
as the current in those strings drops. The increased
current could also cause further LED failures in the
defective string as the current increases. An open
LED will cause the whole string to cease operating.
This will increase current in the remaining strings
by a factor of 1/(s-1) where s is the number of LED
strings connected.
Matrix Connection
To help improve the reliability of parallel connections, the matrix connection shown in Figure
14 may be used. Also called a cross connection
circuit, connections have been added between
the parallel LED strings. This essentially results
in a number of series connected LEDs that are
“stacked” in parallel.
In this circuit, the required string voltage and
driver output current remains the same as our
parallel example. Like the parallel circuit, the
number of LEDs that can be powered without
exceeding the upper voltage limit of the driver is
much higher than with a series connection.
However, the matrix connection, is somewhat more
fault tolerant, and since the balancing resistors
needed for parallel operation are not used here,
the efficiency of this connection is improved.
But even current distribution across the matrix
remains a problem. Inequalities in current flow
(again caused by component tolerances) may cause
visible differences in the brightness or tone of the
light output. Any differences in thermal characteristics caused by current variations could cause
these issues to deteriorate over time.
A shorted LED will cause the parallel row containing
the faulty device to fail, but the remaining rows
will continue to operate normally. If an LED fails
open, only the remaining LED’s on that row will
see an increase in current (by a factor of 1/(L-1),
where L equals the number of LED’s in that string).
Again, the remaining LED’s will operate normally.
MultiChannel Connection
It is typically recommended that series connections
be used whenever possible. This would avoid the
current and thermal distribution issues of parallel
and matrix connections. The most robust connection would be to use a separate driver for each
LED string (or a multichannel driver). This would
combine the control and reliability advantages of
a serial connection with the increased capacity
of the parallel/matrix connections. The obvious
disadvantage to either of these approaches is the
increase in cost and complexity.
Circuit Protection
LEDs are highly reliable devices, with average life
spans that approach 50,000 hours. By far, the most
common field failure is the gradual degradation of
light output to 50% of rated value.
However, failures due to mechanical/temperature
stress, misapplication, faulty packaging, etc do occur. The most common “catastrophic” field failure
is for the LED to fail open. When this happens,
as we have seen, it can quickly cause the entire
circuit to fail.
A common cause of catastrophic failure is the application of excessive forward voltage or current.
The use of a constant current buck regulator (as
shown in our examples) will protect against most
instances of this. However, components can be
misadjusted or surges may be induced by external
circuits or events.
Figure 8 shows protection devices (PDx) connected
in parallel with each LED. Available from a number
of vendors, these devices are typically a form of
voltage triggered switch that activates if the LED
fails open. They then provide a current bypass that
prevents the failure of the rest of the LED string.
Once the LED is replaced, the PD would automatically reset and again present a high impedance to
the current flow. To keep cost down, it is typically
possible to connect one PD across two LEDs.
LED lamps and systems are covered by a variety
of industry packaging, test and safety standards.
Which approvals are important to a particular
project is highly dependent upon the application.
The more common UL standards applied are:
UL 60950
UL 8750
Safety of IT Equipment
(Commonly used with AC input power supplies)
Safety of LED Equipment
(Covers drivers, controllers, arrays and modules)
UL 1310
Safety of Indoor & Outdoor
Class II Power Supplies
UL 1310
Safety of Components for use in
Signs & Outline Lighting Systems
There are similar
i il (or
( even harmonized)
i d) standards
t d d
or norms issued by other agencies. At this time,
no single standard has emerged as an industry or
market requirement for the driver portion of LED
International Protection (IP) ratings are typically
used as criteria for LED driver packaging. Defined
in IEC 60529, they classify the level of protection
provided against the intrusion of solid objects,
dust, and liquids by electrical packaging. It consists of the letters IP followed by two digits. The
first digit indicates the protection level against
the ingress of solid objects, and the second digit
against the ingress of water.
Most LED drivers are rated at IP67 or IP65. The first
digit (6) rates the package as totally protected
against dust. The second digit (5) rates the package as protected against low pressure water jets
(limited ingress permitted), while (7) rates the
package as totally immersible.
In Summary
The use of LEDs is growing at a very fact pace. This
rapid expansion in their application is driven by
maturing technologies that have increased their
cost effectiveness and the introduction of many
new products.
With this note, we have tried to provide an overall
look at the issues involved in powering LEDs. When
applying LED drivers use the technical expertise
of your vendor.
Web: www.micropowerdirect.com • Email: [email protected] • Tel: (781) 344 - 8226 • Fax: (781) 344 - 8481
Page 7
Go Direct
For Low Cost, High Performance LED Drivers
Wide Input DC/DC LED Drivers
Miniature SMT DC/DC LED Drivers
Compact, Wired DC/DC LED Drivers
• Up to 1A Output
• Constant Current Output
• 7V to 60V Input
• Constant Current Output
• SMT Package
• 5.5V to 48V Input
• Constant Current Output
• 7V to 30V Input
• IP67 Rated
• Efficiency to 97%
• PWM & Analog Dimming
• PWM & Analog Dimming
• Up to 700 mA Output
• Efficiency to 95%
• 4” Wire Connections
Very Low Cost DC/DC LED Drivers
Compact, 0.3 - 1A DC/DC LED Drivers
• Constant Current Output
• Efficiency to 95%
• 5.5V to 36V Input
• Compact Case
• Constant Current Output
• Up to 1A Output
• Efficiency to 95%
• PWM & Analog Dimming
• Up to 700 mA Output
• Wide Input Range
• PWM & Analog Dimming
Application Notes
Download Free From MPD Web Site
Go to MicroPower Direct for all of your power requirements. With well over 4,000 standard products, we have the product, performance and price that’s a good fit for your application. Call today for very quick service and a surprisingly low price quotation. Or go to:
292 Page Street, Ste D Stoughton, MA 02072 • TEL: (781) 344-8226 • FAX: (781) 344-8481 • E-Mail: [email protected]