What is the hardest way to drive a LED with a single Li-ion cell?
AFAIK, the cell voltage under load must match the LED Forward Voltage (Vf) under current. Suppose I have a fully charged cell (4.2v open circuit) and connect it to a Cree LED. The XR-E datasheet (http://www.cree.com/products/pdf/xlamp7090xr-e.pdf, top of page 7) tells me that the LED's Vf is approx. 3.5v at 600mA and 3.8v at 800mA. So the battery voltage will drop (as the current increases) due to internal resistance until it matches the increasing Vf of the LED.
The question - can a driver find a better (meaning more light) equilibrium? How?
What happens is if you apply more voltage, the LED draws more current. So on direct drive, it really draws a lot of current because of the high voltage (though there is always some voltage sag under load and really a 4.2V cell at rest will be more like 3.8 or 3.9V under load and you can only get so much current out of a battery). The linear regulator drivers prevent the LED from getting any more than a certain amount of current and the excess voltage is converted to heat (I think), but this is still pretty efficient since there isn't that much voltage to burn off.
A lot of budget drivers are really just direct drive, but they apply modes using PWM. Then some have regulators that keep constant output and some others are bucking or boosting, but again with the goal of providing some set amount of voltage and/or current.
If I understood correctly, the answer would be no - direct drive is as high as the LED will go on a single li-ion cell.
But my question is - is there any way to boost the voltage and somehow crank the LED up a notch? Another way of putting it - how can I fry a properly heat-sinked XR-E using a single li-ion cell?
(I know I can fry it storing very high voltage in capacitors like the mosquito rackets do... The "fry led" question should use semi constant current...)
Heat is the enemy of semiconductor devices. LEDs are semiconductor devices. Any excess power is going to show up as heat. The LED is not 100% (or even close) efficient. The best LEDs may use 30% of the energy going into them to create light, the rest is heat. Get the LED guts up to 100oC and it is going to die.
You can drive an XR-E at 1.4 amps with adequate heat sinking - but the LED life is going to be short. Using high current capable cells (low internal resistance) will help.
The LED will fry with too high a voltage - I cooked several with a 7.2V supply capable of about 150A. This will destroy any white LED connected to it.
Best bet is to ramp up the voltage and measure the current to the LED. Stop increasing voltage by 4.5-5V or when the current hits 1.4A - any more is just going to kill the LED. Stick the LED to an old PC processor heatsink - preferably with a fan on it to keep the LED cool.
I think SPAMBOT understood better what wanted - a very bright flashlight with a single (as dangerous as falling fish in the sahara...) li-ion cell. I'm very sorry but English is not my native language.I don't want to fry it, just to make it as bright as possible and understand the differences that make some flashlights better than others.
Please correct me if I got it wrong, but so far, what I understood is:
- You have to heatsink the LED. Heat is the enemy and the risk. It will kill the LED faster than any other thing.
- If you're using a single battery you're safe.
- If you have a boost/step-up regulator capable of 1.5v only and use 14500 you will fry the regulator, not the LED.
Back to theoretical (bot this time more specific scenario) questions:
Suppose I'm willing to shorten the life of the LED 100 times (it will still last longer than the battery powering the torch) and drive it as hard as I can with a 14500. I'll be using a relatively large body (MyLite's J4) and thermal paste, so the LED will be reasonably well heatsinked...
The question comes to this:
- If I use an incorrect step up driver, can I use a driver powerful enough to fry the LED? Or will the driver always be the weak link?
- Can a single 14500 cell provide enough energy to fry the LED?
- The more complicated one: Is there any sense in using a step-up driver since the open voltage of the battery is higher than the initial Vf of the LED?
First Q: Whichever part has the lowest maximum power (or current) rating will be the limiting factor. It may be either the LED, the battery or the regulator. Using a high current step up converter with a low current LED will fry the LED if the battery is able to supply enough power. The regulator will only be the weakest link if it is the lowest rated component.
Second Q: Yes but it will push the 14500 to its limits. By "the LED" I assume you mean XR-E or equivalent.
Third Q: Yes. The cell voltage will drop as the battery depletes but a step up converter can help even out the discharge curve. However a step up/down ("buck/boost") regulator is better because it does not go into direct drive when supply voltage is greater than forward voltage.
The answer is simple. If you want the brightest from an xr-e, the proper way is to get a 1.5A xp-g driver. These only exist as buck driver. Or just get a xp-g with the same thing. It'll increase the temp tint of your 6500K WC to 8000K+, but whatever.
Direct draw depends on pure luck as Vf of each emitter varies. The brightness decreases as the battery depletes, which kind of ruins the point of the exercise. It's only used as a low/zero cost workaround.
In the "real world" they act just as described above. The emitters are mostly like simple diodes. They're specced up to to a certain current but they'll run above that with reduced lifespan. 50% over will probably work and lifespan hardly matters for flashlights.
For tint change, just use google. Prolly get some link to cpf. In any case, it's well known. So get a tint a couple bins below what you want the color to turn out after "overclocking".
the problem with budget stuff is that you don't get trustworthy specs. I learned about electronic circuits with "ideal" equations where if you connect a battery and a diode you get infinite current.
So now I have to add lots of "internal resistances" to the model and I cannot get data sheets I trust (Hey, I can't even get the correct LED sometimes... ).To make matters more complex nothing is linear. For me this is part of the fun but it is sooo much harder to understand the final circuit.
This is an old thread.
I have a problem. I fried my MF UF 3mode driver while soldering it to a diy pill or attaching a new spring with too much heat to it. Now it only has one mode, i think it is high mode.
Where do i get a 17mm direct drive driver on high with 3 modes like the UF one? I need direct drive because i want turbo mode for my U3 led.
but it is sooo much harder to understand the final circuit.