circuits, again using the MC34063, are designed to drive ultrabright
LEDs at constant current. Two suggestions have been made by other
some LEDs may have very limited reverse voltage and so it may be
worthwhile protecting them with a reverse biassed diode or zener placed
Use the current limiting feature of the MC34063 and ignore the control loop.
Use the control loop with the LEDs in the feedback circuit.
circuit is very simple but not particularly accurate. The MC34063 will
limit current when the voltage across the Vcc-Ipk pins reaches a value
between 0.25V and 0.35V. This also varies with temperature. Rlimit can
be set accordingly. While the LED doesn't require a dc voltage, the
output filter is still necessary to smooth the current flow for the
limiter to work effectively.
circuit promises greater accuracy, if that is a requirement, as well as
more flexibility in choosing the current setting sense resistor value.
Another advantage is that the power loss in the sense resistor can be
reduced for both heating and power efficiency gains in higher power
applications. The disadvantage of course is a more complex circuit.
trigger voltage at the comparator input ranges between 1.225V and
1.275V. Without the amplifier the power loss in the sense resistor
would be excessive. The amplifier gain can be chosen so that the
voltage drop across the sense resistor is kept very small, say 100mV or
less. Care will need to be taken when high gains are used, as
interference and noise will be amplified.
an example, consider driving a series of high intensity starleds at 1W
each. The voltage across each of these is approximately 3.3V so about
350mA current is needed. Any number of LEDs can be connected in series
the total voltage across them is within the power supply range (less
that across any intervening components). If we
choose a 1 ohm, half watt R5 then it will dissipate about 0.12W and we
will need to
amplify the 0.35V across it to 1.25V, i.e a gain of 3.57. An R4 of 3.6K
and R2 of 10K will give a gain of 3.8. R3 must be omitted. Measurements showed that the circuit drew 1.8W, an efficiency of 70%.
The circuit can provide a variable intensity by making R2 to be a
variable resistor. If R2 is linear then the luminous flux (perceived
light intensity) will be approximately linearly variable with the resistor's shaft position.