Looking for advice on driving 3w and 1w LED's

I am going to install some high power LED's on an FPV Airplane for nightflying purposes. I want to drive 5 3w LED's, and 2 1w LED's with a minimal of power usage as possible, so I think PWM would be the best bet also so I can control brightness. The 3w LED's (White), have something like a 3.2v forward voltage and 750ma current draw at full brightness, the other 1w LED's have similar forward voltage but 350ma current.

I am considering the ULN2803 but I am concerned with heat dissipation as well as PWM ability, so what does everyone recommend? Ultimately I will probably only run these lights 1/4th of their power if not less most of the time except for take-off's and landings.

What voltage is the battery to supply the LEDs?

I am considering the ULN2803

Don't.
For driving high power LEDs you need a constant current supply, resistors do not cut it at these powers.
To maximize efficiency put the LEDs in series to match the volts drop to as much as possible to your supply voltage.

Depending on your battery voltage a buck or boost constant-current drive module would be the most efficient - only fairly slow PWM works with these from my researches, but the Arduino's default PWM is quite slow.

If the battery voltage is 5V you can use a non-switching constant-current drive chip (I've just been looking at the various CAT4101 chips for instance that do this. PWM fast or slow is fine. 5V to 3.2V will waste some power, but the driver is probably cheaper and lighter. Some heatsinking needed though...

For driving several LEDs in series you might need a boost driver, but otherwise a buck driver likely to be most efficient.

My main battery will be a 4 Sell lipo, so battery voltage will be between 14.8 to 16.8 (as the battery drains), though I am open to putting in a separate 3 sell (11.1 to 12.6v).

The layout I am planning is 3 of the 3w's upfront to be a headlight, so they can be pwm'd as a group and would be ideal for a series, 2 of them will be mounted on either side of the fuse to light up the wings as well as provide flood and landing light, they could be series too, then there will be 1 1w on each wing tip (red/green).

I have been thinking about all of these factors and I know that any sort of linear DC-DC is going to dissipate a LOT of wasted heat and considering the plane is made of foam excess heat is not preferred :).

Something like this but more current and cheaper? 7027 "BuckToot" LED Driver Module - COM-09642 - SparkFun Electronics

Couldn't find a breakout for the ZLED7001 chip (one of the many suitable buck chips - needs support circuitry to work). If want many LEDs all switched then a single DC-DC converter makes more sense to supply them I think.

What about STCS2ASPR? http://www.mouser.com/ProductDetail/STMicroelectronics/STCS2ASPR/?qs=sGAEpiMZZMsE420DPIasPoxeLwn8SKwbvNs8EYaD8YU%3D

Rated to 2 amps constant current sourcing... and can handle the voltage to run some in series plus has PWM pin :).

I also found the cat4101 1a

I ended up ordering some NCP3065PG and CAT4101, I will give status update on them once they arrive.

Does anyone have an eagle part for the STCS2ASPR? If you could point me in the direction even for a PowerSO-10 package that would also be very helpful.

brs0906:
Does anyone have an eagle part for the STCS2ASPR? If you could point me in the direction even for a PowerSO-10 package that would also be very helpful.

Or consider making an Eagle library yourself - its not as hard as you might think (tedious, perhaps, hard: no) - there are lots of Eagle
tutorials out there to step you through it. You can copy packages and symbols from other libraries (alas that is a bit involved, find a tutorial) too.

Ok sofar I am working with the CAT4101 1A constant current sink, for my white LEDs I set the current to 700ma and it runs quite well and use arduino to handle PWM just fine. Now since my design is using a 5v rail and all of these LED's I am using have FV of roughly 3.5 to 3.8v the driver doesn't have to waste much energy on voltage drop so it stays cool, and the led's get warm of course but not unmanageably so. Now what I am considering to reduce component count is to power 2 1 watt LED's in parallel off of one CAT4101, they both can do 350ma so the driver set to 700ma should have no problems handling both in parallel, however since the red LED has a FV of 2.2v and not 3.8 like green, will this cause a current split difference between red and green, and if so how much and how can I calculate that and correct for it?

MarkT:
Depending on your battery voltage a buck or boost constant-current drive module would be the most efficient - only fairly slow PWM works with these from my researches, but the Arduino's default PWM is quite slow.

Most of the controllers for these sort of LEDs have PWM inputs on the board so it's not a problem.

MatCat:
I have been thinking about all of these factors and I know that any sort of linear DC-DC is going to dissipate a LOT of wasted heat and considering the plane is made of foam excess heat is not preferred :).

At least you won't have a problem with airflow... :slight_smile:

fungus:

MatCat:
I have been thinking about all of these factors and I know that any sort of linear DC-DC is going to dissipate a LOT of wasted heat and considering the plane is made of foam excess heat is not preferred :).

At least you won't have a problem with airflow... :slight_smile:

Actually no I won't ROFL, I just did a mounting test as pictured below, zip-ties go through a flattened copper tube to hold the LED star on which makes direct contact with the copper, this is in turn zip-tied to a wooden dowel, which is hot-glued to a piece of foam. Now the LED star stabalizes at 160F when running at full blast (350ma), the back end of the copper as shown get's up to about 140F, and between the wooden dowel and the foam 111F, where the hot glue is doesn't even get above ambient, so I think all is fine.

But my question remains from above, can I drive both a red (2.2v FV), and a green (3.8v FV) which both have 350ma operating capacity from a 700ma constant current sink and have them equally split the current (5v source), or will I need a resistor on the red to keep it from sucking up more current then green?

I think I missed the bit when we went from 10v battery to 5 v supply, but you could put the red and green in series ( 6v ) and run them from a pwm constant current current device from the main battery?

Actually the primary system battery is roughly 15v, that would be a lot of wasted heat to linearly lose in voltage drop! So instead I am going to power the lighting system with a switching regulator that can provide 4A at about 5.2v. Series would be preferred but the voltage isn't quite enough, I don't know if the UBEC's I have coming are switchable to 6v, if it is I will just run those two in series, there is a .5v drop on the current sink too to keep thought of, that puts the theoretical need at 6.5v, however the green LED is spec'd down to 3.4v, so 3.4v + 2.2v + 0.5v = 6.1, and if the UBEC has 6v selector realistic output will be 6.2, I actually have another 7A UBEC that puts out 6.2, I might just use that one.

Oh another thought on these red and green LED's... The actual LED's will be each on about 1.2 meters of 22 AWG wire, which means in series from + to - there is a travel of about 4.8 meters of 22AWG wire, I wonder what that does to resistance and capacitance, and if it is enough to merit bumping up the total constant current sink to overcome the loss, plus it will all be on PWM so technically it should be able to be a bit higher because of that too, any thoughts?

I would not bother about the length it is short enough.
Mind you it will radiate like stink but then you are not going for FCC approval are you.

MatCat:
there is a travel of about 4.8 meters of 22AWG wire, I wonder what that does to resistance and capacitance

At these power levels it might drop half a volt along the wire because of resistance. No biggie, but make sure you have enough overhead in the power supply.

MatCat:
and if it is enough to merit bumping up the total constant current sink to overcome the loss

All you need to worry about is that is has enough volts to play with.