Most of my parts want a 3.3V voltage supply. Most of my projects relay on a single coin cell battery at 3V. But driving an RGB LED from 3V is no fun at all. Lets take for example a nice strong RGB LED like the Osram LRTB G6TG – the green channel runs at 3.2V, the red channel runs at 3.7V – much more than a coin cell can provide. You have to drive them with a higher voltage.
To drive a RGB LED from 3V I designed a breakout board with an integrated 5V step up converter.
Schematics
To try the feasibility of an low power LED I have built‚ little breakout board to drive them with the MAX619. It is an Regulated 5V Charge Pump DC-DC Converter – you just add some capacitors and get an regulated 5V output. Using the schematics from the data sheet, adding the RGB LED and using some BC850 as switching transistor you get something like this:
The PCB
Laying out all this stuff on a tiny package was – again – not easy. In the end I put all the step up circuit on the bottom:
And the LED and the transistors on the top:
Testing it
Putting all this on a breadboard you quite a strong light, which can be controlled by any PWM signal (via the CTRL pins):
Further Options
The same circuit works with an Osram LRTB G6SG – which has just 3 separate white LEDs – this will give you quite a bang of white led power. You can control the three LEDs separately which gives you an three times bigger PWM resolution for free (which is just an euphemism for ‘you would design a different board but I am just too lazy‘).
The LRTB G6TG can handle about 50mA per channel and 400mW total power. But this can’t be delivered by the MAX619 (it has a 50mA maximum). A real high power solution would use something like a MAX682/MAX683 to put out 5V at 250mA – enough to blind you.
So if there is enough interest I will design a board, which perhaps become a kit in the near future – put your preorders in the comments.
{ 2 trackbacks }
{ 3 comments… read them below or add one }
The Sharp GM1WA55321A is tiny, but bright (for its size) and uses 3 volts or less for each LED. 5mA, so it isn’t going to be a flashlight LED, but for indicator purposes it works well.
Oh thanks, this is a very interesting hint.
I think driving LEDs for indication purpose from 3V can be done.
But I often want to illuminate something form a coin cell – and this can be a major issue
If you have to do variable color illumination it probably wouldn’t work, but there are plenty of really bright white LEDs
I think I have a Seoul Semi P7 – 900 lumens at 3.6v, but is still very bright from a coin cell. I got it from DealXtreme – they also have similar IR Leds.
Speaking of which, digikey.com and perhaps others now has the ATtiny45 in an TSSOP package. 1/3 to 1/4 the size of the SSOP used for the uTVBG. I’m prototyping using larger device but have some of these. I thought it was a mistake but Atmel confirmed they do make them and are correcting the datasheet and other informaiton:
http://www.atmel.com/dyn/products/packagecard.asp?family_id=647&part_id=2493&package_id=319&green=0