AA batteries v LEDs

[Cursitor Doom]

Tabitha has got her quoting mixed up by the look of it; I for one couldn't make head nor tail of it.

[tabbypurr]

nope

no surprise.

[Martin Brown]

Depends what you mean by "turn on". White LEDs at 5uA typically drop around 2.4v and the die is just visibly glowing at that. Dark adapted you might see a glimmer at much lower currents with the latest crop of high efficiency LEDs.

There are quantum limitations on the terminal voltage having to be >1.25v to actually emit a blue photon! Usually I find at least 2v.

But a lot of modern led torches now use a voltage to current converter chip that takes the battery voltage from a single cell and converts it to an approximately constant current drive. Outdoor solar powered nightlights use a similar method to drive their LEDs too.

The result is that the LED will be just as bright but for a shorter period of time since the device cuts off to protect the battery from corrosion induced leakage at some minimum terminal voltage.

That internal resistance might matter if the LED was drawing several amps like a torch with a bright filament bulb does.

[Cursitor Doom]

Still pissed from xmas, NP?

If you'd been sober enough to read and understand the thread properly you would realise I'm referencing small signal diodes here (0.55-0.72V BH) for the sake of simplicity in making this comparison - NOT coloured LEDs (exact same principle but an easier range of values to deal with).

[Cursitor Doom]

I can't recall if it was you or someone else who asserted that a constant voltage across an LED gives rise to a constant current through it. Diodes, including LEDs, DO obey Ohms law, of course, but are NOT linear loads like resistors. The barrier height of these components is temperature dependent. When you apply a voltage to a diode, the resulting current generates a heating effect which lowers the barrier height causing more current to flow which causes more heating and so on. It's a vicious circle which often leads to the eventual destruction of the device as it generates more heat than it can dissipate. This can be avoided by using a current-limiting resistor in series with the LED calculated to provide the specified forward current the manufacturer recommends. That's the quick and dirty way of doing it, and it works well enough in most circumstances, but the proper way to supply LEDs, though, is from a properly regulated current source. Hope that's finally cleared that up!

[newshound]

I'm a bit confused here. It's been ages since I looked at any quantum theory and ICBA to do it now, but Martin is normally pretty fussy about getting the science right, so I'd be surprised if he is wrong here.

I have a little Cree torch here which produces pretty white light. It has a yellow die, presumably a phosphor triggered by short wavelengths. It also runs on a single AA cell, including an Eneloop.

Doesn't that mean there must be a voltage multiplier circuit? In which case the type of battery or its voltage is irrelevent.

[Gunther Heiko Hagen]

The most elementary and cheapest form of voltage multiplier is a ladder- like network of capacitors and diodes. Can you see anything like that inside the torch?

[Chris Green]

I think that's quite likely though I don't really know. As you say, if there's electronics between the battery/cell and the LED then anything *might* happen with different cell technologies.

[Chris]

Better look for some sort of oscillator as well; those multipliers won't run straight off a battery.

[The Natural Philosopher]

And once you have an oscillator, its a toss up as to which is cheaper - a chain of diodes and capacitors or a little bit of wound ferrite.

[James Wilkinson Sword]

Wrong. More than a volt.