What is the best LED driver available for Li Ion battery?

I wanted to drive some discrete white lower power LEDs that range in power from 100-300 mA. Is there a pre-made and SMALL driver where I could just connect the battery and LED and have it work at the correct current and voltage (3V- 4.2V)? I have seen some around but most are HUGE. I would like one as small as possible. I know they exist because every maker of flashlights use them with their LEDs and they are small enough to put in their flashlights.

Does anyone have any information or recommendations?

Thanks

There's a lot to choose from at Mountain Elecrinics.
I you could start there........

http://www.mtnelectronics.com/index.php?route=product/category&path=67
P

I am not 100% sure what your goal is. Exactly how are you planning to wire them? number of leds, power source, intended use, and so on.

IF you want budget stuff, look up fasttech, gear best and banggood. I get a lot of drivers from there. Of course if I need speed, quality and service Mtn Electronics is my goto. Search Led Driver

I just want to drive some 5mm low power Jeled HK LEDs (about 100 mA- 300mA) with a single li-ion battery. No modes or other frills, just a simple and small driver. Those ones at Mountain are for the bigger emitters.

5 ohm resistor in series with the battery and LED. That is as simple and reliable as it gets.

Here's the logic:
- The battery voltage (V battery) is only slightly higher than the LED forward voltage (Vf)
- Every driver needs at least some amount of voltage coming from the battery that is higher than the LED Vf (expressed as Vbattery - Vf). If this number is less than 1 volt, it gets more complicated, and that is the range you are in with a 1 cell and 1 LED flashlight.
- The "efficiency" of a driver largely determines the battery life, and in this region of fairly small (Vbat - Vf ) a resistor is as good as 90% of the drivers out there.
- In the 50 - 500 ma range, most batteries will hold their voltage under load fairly constantly, so you can, to some extent, ignore that this is changing some during discharge.

Calculation used to determine resistor properties:

First, you estimate the typical Vbattery and Vf under load:
- Vbat = about 3.8 volts
- LED Vf = about 3.3 volts

and the likely worst case scenario
- Vbat = 4.0 volts
- Vf = 3.0 volts

Target 100 ma current, max 300 ma

From ohms law, V = IR, so (3.8-3.3) volts = (0.100 amps ) x Resistance, 0.5 volts = (0.100 amps) x R. R = (0.5 ) / (0.100) = 5 ohms

Now do a worst case

(4 - 3) volts = (0.100 amps) x R, 1 volt / (0.100 amps) = R = 10 ohms.

More or less, you can be sure that a 10 ohm resistor is the maximum value required, and a 5 ohm is probably sufficient. There are other resistances in your light, so don't worry too much about being exact.

Finally, calculate the power rating of the resistor:

Typical Power = R x I x I, so Power = 5 ohms x (0.100 amps ) x ( 0.100 amps) = 0.050 watts.

Power worst case = 10 ohms x (0.300 amps) x (0.300 amps) = 0.9 watts.

This means you should buy something like a 1/4 or 1/2 watt rated resistor and provide a thermal path for it to the flashlight body.

I thought about a resistor but I would just be wasting power in the battery wouldn't I? Li ion batteries can be as high as 4.2V on full charge and drop to 3V before they cut out if protected. This means if the LED needs 3.3v I would need to be able to drop .9v on the resistor to be safe. This means The LED would go dim when there is almost a useable volt left in the battery. That seems like while it would work it would reduce run time by allot. If I want to use these small LEDs in a keychain type light the battery is small too (like a 10180 or coin cell).

Don't I need some sort of buck-boost circuit? It will buck power over 3.3v then boost it when it drops below that? How is it done in the industry? Do they use resistors or some sort of constant current IC?

You are right, that a fully charged, sitting on your desk battery is around 4.1 - 4.2 volts. The trick is what happens when you start pulling power from it, and very quickly, the voltage drops to around 3.8 volts.

Take a look at this cpf thread on testing Li Ion cells: http://www.candlepowerforums.com/vb/showthread.php?117117-Li-Ion-Battery-Shoot-Out

While the exact curve shape depends on the cell construction and load, the general shape of these curves is similar for most cells. Almost as soon as you turn it on, the cell voltage will drop to around 3.8V or lower, and by the time the Vbat is in the 3.5 - 3.4 V range (under load), there is not much left. Consuming the cell down to 3.0 volts will rather dramatically shorten its cycle life.

I have built some hot-rod lights that pushed cells well beyond their design, but I also threw away the cells every 5 cycles for safety precaution reasons.

Sorry for the short info in each posts, for some reason I get glitches when posting sometimes.

If you assume that the "really useful" part of the battery voltage is between 3.8 and 3.5 volts, and the LED Vf is a constant 3.3 V (which it actually isn't), then the worst case loss is (3.8 - 3.3) or 0.5 volt.

Assuming that we ignore losses like springs, contact resistance, etc, and assign all losses to the load resistor, then the worst case power lost to the resistor is (volts) x (amps) = (0.5 volts) x (0.100 amps) = 0.05 watts.

Compare this to the power used in the LED of (Vf) x (current) = (3.3) x (0.100) = .33 watts.

Using lost efficiency as (resistor losses) / (total power) = (0.05) / (0.05 + 0.33) = 10 %. That is not that bad. It takes a seriously well designed converter to be at only 5% losses. Some of the better drivers on the market need at least 0.2 volts of "headroom" over the LED Vf to operate, and then provide 90-95% efficiency, starting from this 0.2 volt head room loss.

I have built lights with drivers, and I own some lights with drivers, but after 10 years, the most reliable ones are the ones that I built using resistors. A lot of it has to do with intended use.

If you have a light where you are pushing the LEDs very close or beyond their rated current limit, then very good thermal paths and fairly precise control of the current is needed. If you are selling lights, the market is not particularly kind to resistor based current control, mostly because people's expectations are very high vs their understanding of what is possible.

For a small, single cell light, for modest output, personal use, a resistor based setup might work.

The nice thing about the LED flashlight hobby is that you don't need to build up the full light to see how well it will work. Everything can be contained in a drawer of small box. For example, I use those clear boxes with purple handles from Target (less than $10) to keep a "hobby in a drawer". All of the parts you need can be purchased from Mouser, Digikey, or Newark. I tend to use taskled drivers for serious projects, but there are lots of driver suppliers. (not so many at 100ma)

Nice to see someone else building lights without drivers, Harry.
P

Thanks Peter.

For DaveC1964, in the electronics section of the forum, there are 'threads of interest",

http://www.candlepowerforums.com/vb...ronics-Batteries-Included-Threads-of-Interest

one of them is on drivers.

http://www.candlepowerforums.com/vb/showthread.php?188227-Driver-board-list-regulator-board-list

For the current levels that you are targeting, and very small voltage drop, it will take a careful sorting through. The good news, is that compared to the past, all of this stuff is inexpensive.

Thanks Harry for the links. It seems most of those are for high power LEDs (at least higher than 100mA)unless I missed one.

Booster circuits don't work if the voltage is higher than the boost volts, which in the case of a li-ion it would be. Buck circuits would quit too soon. It may sound dumb but how about combining a booster circuit or joule thief and a resistor? The resistor would be at the battery before the booster and drop voltage to an acceptable range for the LED and the boost circuit that outputs 3.3V would keep it there for as long as there is energy in the battery. Has anyone tried anything like that and would that work better than just a resistor? This way the LED wouldn't quit with almost a volt still in the battery unused.

Where is this "almost a volt" that you are talking about ? At full charge, under load, the battery voltage will only be 3.8 volts within 10 seconds, so it only really starts with 0.5 volt differential. The other part that is not so obvious, is that the LED Vf is not a constant. It might take 3.3 volts to push 100 ma through one, but at 3.0 volts, it might still pass 50 - 75 ma, so it will go through a slow dimming, vs on / off failure with a low battery.

Try building up a proto on a piece of plywood (forget the flashlight body for now). Just mock up a battery, wiring, LED, and a few versions of drivers. Take some measurements with a volt meter around the circuit to see where your current really ends up at, and if you are happy with the setup.

The great thing about playing with LEDs is that they are so cheap, especially now. When I started playing with them, we were paying $50 / each for LEDs that produced under 200 lumens. I burned out more than one 'learning" about LEDs, now that was painful. Today a similar LED is around $1 /each. Its a hobby, time to get your feet wet and have some fun.

I am serious about one thing though - buy some sun glasses and consider to have a cardboard box around to help block some of the light.

There are a lot of "1W" drivers around, and many higher power ones that could be adapted to 1W. Many of them can be had for a few dolllars, or even less. Anything with an AMC7135 on it is what I'm talking about. They have an input voltage of 3-ish to 5-ish or 6 volts (vendors play fast and loose with the specs), and an output current that's a multiple of 350 or 380 mA.

An AMC7135 is a 350 or 380 mA driver chip. It's great for running white LEDs on single LiIon cells, and it's really cheap. Put 4 of them on a board, and you get a 1.4A driver (4 times 350 mA). With 8 you get 2.8A. See a board you like but the current is too high? Just remove some of the chips (or cut the circuit traces). Another thing you can do if the current is too high is to put several LEDs in parallel to increase the current capability. This usually works pretty well if you have identical LEDs from the same batch.

The driver boards typically include a microprocessor to control 'modes'. They do this by turning the '7135s on and off really fast, and controlling the brightness by controlling the relative on and off times. This is called Pulse Width Modulation, commonly called PWM. Though digital cameras can, your eye can't see the flicker if it's over 100 Hz (often it's 500 Hz or higher because up to 200 Hz or so it gives some people headaches).

If you are unsure which direction to head, just pick up either a Lumileds Rebel or some similar part pre-mounted on a star heat spreader board.

Try a few different drivers at your targeted current of 100 ma. Those LEDs have so much headroom over what your goals are, that further heat spreading is probably not needed, and if the current is accidentally too high, it won't matter.

Here you go;

Single Level, 350ma Constant Current, Li-ion input, Super Small (11mm), Top Quality.

http://theledguy.chainreactionweb.com/product_info.php?cPath=48_49_61&products_id=1342

These were originally built for the Draco keychain flashlight, which used the 10xxx series li-ion rechargeable cells.

-Michael