Jinx wrote: >> It would kill the backlight of the LCD. Read the specs on max >> ripple for the supply and you will see that they have a very low >> tolerance. >> >> I did this once, (by mistake? / ignorance) and the backlight gave up >> after a month! >> >> >> Kyrre > > That doesn't sound right, something else must have been going on. > LED backlighting is simply that - LEDs, and they should be able to > take all the punishment we dish out to normal LEDs from DC to MHz > PWM. It is very important to have a current limiting resistor though > because the LEDs aren't designed to run off 5V directly > It is very difficult to visually notice an overcurrent on LEDs. Increasing from 15mA to 30mA you will see a little increase of brightness, not a visual indication of overcurrent. Increasing more, to 50mA or even 80mA you will see the LED shinning more, but going to 100mA it will start to reduce brightness due the heat generated until it blows up the junction. Some backlights are made with a parallel / serial combination of LEDs, but without even looking for the datasheet, you can guess some numbers; Suppose the whole backlight uses 8 LEDs, each led 1.7V @ 15mA Option 1) ========= 8 LEDs in parallel, it would drain 1.7V @ 120mA. Your 5.1 Ohms resistor would hold 5.1 x 0.12 = 0.6V, what is wrong, since if the leds would hold only 1.7V, R5.1 should hold 3.7V at least. Option 2) ========= 2 blocks of 4 leds, connected in serial. It would hold 3.4V @ 60mA, then R5.1 would hold 5.1 x 0.06 = 0.36V, and it should do 1.6V at least. Option 3) ========= 3 blocks of 3 leds (suppose it uses 9 leds), connected in serial. It would hold 5.1V (5V is ok) @ 45mA, then R5.1 would hold 5.1 x 0.045 = 225mV What seems to be the correct answer. But if 3 Leds in Series needs 5V, why use the resistor? safety! The Voltage across the LED junction grows much slower than current. Applying 1.8V across a LED, the current would go almost twice of would be if applying only 1.7V. So, imagine something going wrong with the voltage regulator, and it applies 5.25V instead 5.0V, that is the stated +-2.5% error from LM340T datasheet. With this 5.25V applied directly to the LEDs, they could increase the current in 50%, what would destroy the LEDS in few months of use. The 5.1 ohms resistor, even holding just 225mV when Vcc=5V, would increase substantially this Vdrop if the LEDs current increase from 45mA to 70mA. 5.1 x 0.07 = 0.357V, it means, the Resistor VDrop went from 225mV to 350mV, what would create a "automatic" strong current limiter without interfering much when VCC is 5V. This extra Vdrop of 125mV will pull down the voltage applied to the LEDs and they goes back to a safe current consume region. When you use a R10 ohms, the 45mA would generate a RVdrop of 450mV, what would reduce VLEDs to around 4.55V, current will drop substantially to reduce RVdrop and create an intermediate LEDsCurrent and bright. At this very short range of Voltage (less than 500mV) is somehow more difficult to control the brightness you want (day/night). What I would do it exactly keeping two 5 ohms resistor in series, and short circuiting one during day-time. The resistors are important to the current limiting job. +5V o | .-------o-----o | | | R1k | | | |uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads