Question about LEDs and battery power

[mithrandiir42]

First some background...
I'm working on a project that consists of 20 rgb leds powered by 2 900ma 3.7v lithium ion batteries.
each color per led needs ~20ma of current
so my project needs at least 20*3*20 = 1.2Amps if all colors of all leds were full on.

for the sake of simplicity lets just assume that each color (in reality each color has slightly different tolerances)has a max voltage tolerance of 3.8v (in reality each color has slightly different tolerances). I'm using npn transistors to control the on/off state of each color. now my problem is that I want to get full brightness from my leds and since there's roughly a 0.7 voltage drop across the transistor, the max voltage my leds could get it 3.1v which doesn't produce full brightness.

Now to the real question.
What is the best solution in terms of efficiency (longest run-time)?
put two 3.7v batteries in series to get 7.4V and then use a LDO voltage regulator such as an lm317 to output 4.5v (3.8+.7) so that my leds can receive the full 3.8 volts when the transistor is full on? What happens to the extra voltage here? is it wasted? or is it somehow put toward the load giving me more than 900ma since i'm only using little more than half the voltage?

...or place the batteries in parallel so i'd have double the current 1800ma
and find some kind of DC boost converter solution to bump my voltage from 3.7v up to 4.5v and be able to supply at least 1.2Amps? obviously this would pull more current to produce the voltage needed but that's ok.

so far I've been unable to find a module small enough for my application that can deliver that much current. If there is some kind of module that can do it. it needs to be less than .625 inches wide/tall, length isn't an issue

I've been using these in the past when i had more space to work with and they worked great:
Digi-Key - 507-1379-5-ND (Manufacturer - VRAH-02C50B0)


Please pardon my lack of accurate terminology.. I'm a software guy getting into hardware

Thanks in advance for any guidance.

 

[3lectrokid]

an lm317 dissipates a lot of heat so would not be very efficient

 

[mithrandiir42]

Maybe a Buck converter would be better to drop 7.4v down to 4.5? I assume this would be more efficient?

 

[Vizier87]

an lm317 dissipates a lot of heat so would not be very efficient

No they don't if your supply is only 7.4V from Li-ion batteries.

Maybe a Buck converter would be better to drop 7.4v down to 4.5? I assume this would be more efficient?

Buck converters are MUCH better, but they're costly too.

 

[Diver300]

A few points.

The battery rating is mAh, so the two batteries in parallel will give 1800 mA

The forward voltage of an LED is about 1.8 - 2 V for a red LED. Green ones are about 2.2 V and blue ones are around 3.6 V.

A single Li-Ion battery will drop to around 3.4 V when fully discharged and rise to around 4.2 V when fully charged. To get a current that doesn't vary too much between charged and flat means that you need some form of current control, but it is quite possible to have a lower voltage drop than 0.7 V.

https://www.fairchildsemi.com/ds/FA/FAN5607.pdf

https://focus.ti.com/lit/ds/symlink/tps61058.pdf

Those ICs use two different ways of driving LEDs relatively efficiently from variable supply voltages.

How are you controlling the LEDs? Have you got a specification for the LEDs?

 

[crutschow]

I'm using npn transistors to control the on/off state of each color. now my problem is that I want to get full brightness from my leds and since there's roughly a 0.7 voltage drop across the transistor, the max voltage my leds could get it 3.1v which doesn't produce full brightness.

A fully on transistor in a common-emitter configuration will only drop about a 0.1V or so from collector to emitter. (The base-emitter voltage will be about 0.7V under these conditions).

 

[mithrandiir42]

How are you controlling the LEDs? Have you got a specification for the LEDs?

These are the LEDs I'm using
LED - RGB Super Flux [RGB] - $1.00 : FunGizmos, make something fun!
RED: Typical: 2 V Max: 2.4V
GREEN: Typical: 3.4 V Max: 3.8V;
BLUE: Typical: 3.4 V Max: 3.8V
DC Forward Current:20mA Lens - Clear (water)
Configuration - Common Anode

obviously the red requires a much lower voltage, I was thinking about having a separate voltage regulator for the red channel.
(one producing just enough voltage to drive the red, and one to drive the green and blue)

I'm using the six PWM channels of an AVR ATmega328P to control the leds brightness via 2x ULN2605B (Quad NPN Darlington Arrays)
https://www.st.com/stonline/products/literature/ds/1535.pdf

I should mention that half of the leds are controlled seperately from the other half, so that effectively I have two strings of 10 RGB leds giving me two different colors at a time. (that's why I need 6 PWM channels and 2 ULN2605s)

Everything works fine running straight off the battery. but now I want to supply those leds with a constant voltage
that will give produce full brightness from full charge until the battery protection kicks in and kills it at 2.5V.


it looks like TPS61059 might just give me enough current to drive my blue & green channels (20ma*2*20)=800ma and then I'd need a second one to drive my red but the only problem is hand soldering a QFN ic could be pretty tricky

 

[Diver300]

obviously the red requires a much lower voltage, I was thinking about having a separate voltage regulator for the red channel.
(one producing just enough voltage to drive the red, and one to drive the green and blue)

but now I want to supply those leds with a constant voltage
that will give produce full brightness from full charge until the battery protection kicks in and kills it at 2.5V.

You should control the current, not the voltage, to an LED. It is never a good idea to parallel LEDs, and certainly not ones of different colours.

If one has a lower voltage, it will take all the current if they are in parallel.

The common connection on the three LEDs makes circuit design more difficult.