Stephen A. Lawrence wrote: > If, instead of an LED, you put a normal diode in parallel with > a resistor, the diode should get warmer and the resistor > cooler, as noise from the resistor is rectified and the > electrons give up the 0.6 volts of junction energy when they > fall down across the junction. LEDs have a somewhat higher > junction voltage than a common P/N silicon diode -- > typically a couple volts IIRC rather than 0.6 volts -- but > it's the same deal: noise in the resistor carries energy > from the resistor to the diode. > > In fact, you might do better to use Schottky diodes; they > have a lower forward voltage drop and should be able to > rectify smaller levels of noise energy. With an LED with > its high forward drop, very few of the noise crests are > going to make it "over the wall", or so I would expect.
That's usually the case-- 1 to 5 Vf range. The LED's effectiveness drops exponentially below Vf, but still emit, as there's no real lower limit, which is why you would need trillions of LED's each connected to a resistor to emit any appreciable light. LED bandwidth is another factor to consider. Given a year or two I'd imagine a company such as IBM could create a FRE device that emitted appreciable "free" light. [snip] Regards, Paul ____________________________________________________________________________________ Be a PS3 game guru. Get your game face on with the latest PS3 news and previews at Yahoo! Games. http://videogames.yahoo.com/platform?platform=120121
