Questions about powering two 3W RGB LEDS using JeeNode, and a 7.4V 2S Lipo battery
First off, sorry for any n00b questions, and I would really appreciate any advice.
I’ve been working on a project where I want to remotely control 2 of these high-power 3W RGB LEDs: http://www.superbrightleds.com/moreinfo/high-powered/vollong-3w-rgb-high-power-led/899/2214/
I want to use a 7.4V Lipo battery, and run around 300-400 mA per color. Would the LED node be the right choice for this project? Will it run OK if I run it off 7.4V instead of 12V?
I’m also not sure about which battery to use. There are some size constraints - This 500mah one is right on the edge: https://www.sparkfun.com/products/10500
Would I be able to charge this LiPo battery from USB, if I add the USB port to the LED node? Has anyone done this before (See http://jeelabs.org/2010/05/29/using-lipo-batteries/)
I already need to redo the PCB layout to make it closer to a square, and I’ll be adding an embedded plug. I would really like to charge the batteries via USB, so maybe I’ll look into modifying the JeeNode USB and adding the parts from the LED node.
Each of the 2 LED colors will be in serial. Max forward voltage for green/blue is 3.8V, so I believe two in serial would have a drop of 7.6V (?). Do I need to worry about that 0.2V extra, or will it be fine?
I’ve also designed a way to mount the LEDs that includes some room to attach mini heatsinks. I think this is necessary because they’re probably going to get super hot.
Sorry for the scattered thoughts, but thanks for your feedback!
Questions about powering two 3W RGB LEDS using JeeNode, and a 7.4V 2S Lipo battery - Added by martynj almost 5 years ago
You have some more design decisions to make on this project. First, note that the LED’s selected are not like the usual LED strips - there are no built in current limiting resistors. You can add your own, but this will kill off the idea of running two diodes in series - there is not enough “headroom” to define reliably the current with a passive resistor and have enough volts left over to turn two Blue diodes on.
IMHO you have a basic choice
for simplicity, use a standard LED node with single diodes/individual dropper resistors (these combinations can be paralleled) and lose some battery power as heat in the resistors, or get deep into modifying the design to have diode string current regulation/limiting. Possible, but perhaps brave for a newcomer. recall that a USB port provides ~5.0 V at ~1 A. You can use this as the basis of a charging circuit, but you need a step up convertor to convince the electrons to flow into a ~7.x V battery, current limiting, end of charge detection and charge time limiting.
For battery charging, this needs to be a separate exercise
RE: Questions about powering two 3W RGB LEDS using JeeNode, and a 7.4V 2S Lipo battery - Added by nathan.f77 almost 5 years ago
Thanks very much for your reply, that was very helpful!
Heres the data sheet for the 3W LED again - the 3.8V max forward voltage is right on the edge, and typical forward voltage is 3.4V. I know the blues and green won’t be as bright as possible at max current, but do you think it will still work OK? I know there’s not a lot of headroom, but I think it should still work. And if I use the JeeNode USB which includes the LiPo charging circuit, then it will automatically cut off when the voltage drops below 3.4V.
So, then I’ve only got 3.7V, and I’m thinking that I may as well control each LED separately, so I’m going to use 6 MOSFETs. In that case, if each transistor is switching up to 500ma, do you think I should use MOSFETs, or some other kind of transistor? I’m looking at the MOSFET plug info, which says “This plug can be operated at 5V as well as 3.3V.”, so I assume that my 3.7V battery will be fine.
Thanks again for your help! I’ll be sure to post a link to my project when it’s ready :)
EDIT: This is great, the ATmega328 has 6 pins that support hardware PWM, so that’s perfect.
RE: Questions about powering two 3W RGB LEDS using JeeNode, and a 7.4V 2S Lipo battery - Added by martynj almost 5 years ago
The main issue to try to design around is that any battery voltage is actually variable, depending on the state of charge. You need to pick at least two typical voltages (I suggest fully charged and 25% remaining - you can google the discharge curves for all standard battery chemistries). Then you need to use the standard ballast resistor formula to calculate the minimum series resistor for each LED that limits the diode current to below the specification sheet maximum. Then chose the higher of the two values - these ensures that the diode current is limited under worst case conditions.
This will be a compromise value, with the intensity changing as the battery voltage changes. You may be able to do some table-driven compensation if you measure the actual battery voltage and vary the PWM accordingly.
Three MOSFET’s are enough since you can put the diode+resistor combination in parallel with another same colour diode+resistor combination i.e. two diodes, two resistors.
MOSFET’s remain a good choice, using a bipolar transistor as the switch has the Vcesat eating into the headroom more than the Rds drop across the MOSFET.