Oooh. I love a good electronic pissing match.
Quote:
Originally Posted by Dsanduril
True, but [current and voltage] are interrelated, and the LEDs themselves are very sensitive to changes in voltage.
|
Irrelevant in application. When the current through the LED is limited (by a resistor or otherwise) to a safe range, you cannot have destructive forward voltage drop. It's semiconductor physics.
Quote:
Originally Posted by Dsanduril
When you use a resistor as the sole current restricting (let's not say current limiting, because they don't) device with a varying input voltage you run the risk of the current rising to the point of destroying the LED. The calculator assumes a steady supply voltage, run an array design in it using 12V, then change it to 14.4V and you'll see that the design resistance increases by about 50%.
|
I think you'll find that the term 'current limiting resistor' is an accepted industry usage. And, upthread, I worked the calculation using 15 v. That's the
safety margin.
Quote:
Originally Posted by Dsanduril
For the 4V LEDs you can try and measure the current (at 4V) and then add a resistor to get the same current at 12V, or you can most likely put three of them in series and run them at 12V (as Atoll suggested). Either way, you run the same risks as discussed earlier; if you size your resistors for 14.4V you will reduce your light output by about 50% at 12V, and if you size them for 12V you will increase the heat in the LEDs by a significant amount at 14.4V and thus reduce their lifespan (by how much being the still open question).
|
4v was the rating of the lil'
power supply, not the LED. I've yet to meet a 4v LED. So that '4v
power supply' tells us not much, except maybe implying there's a resistor in the LED unit that eats up 1v at desired current.
Quote:
Originally Posted by Dsanduril
From a pure efficiency standpoint, the single LED approach isn't it. 3/4 of your energy will be heating the resistor(s) needed to restrict the current. Better to put some of the units in series so that you're getting light rather than heat for the same amount of electricity. The most common method is to take your supply voltage, subtract about 3V, then divide by the LED forward voltage (again about 3V) to get the number of LEDs that are ideal in series and provides the most lighting with the least loss to resistors.
|
A single LED with a resistor, on 12v is still more efficient than the equivalent incandescent bulb, but yes, it's better to put less into heat and more into light. As you already mentioned, the smaller the drop across the resistor, the greater the variance in current for a small supply voltage change. When just doing the single resistor thing, I aim to have about 2/3 to 3/4 of the min available voltage as LED forward drops, then calculate the limiting resistor using the max expected voltage. This seems to strike a reasonable balance between simplicity, efficiency, and relatively steady brightness across the expected supply variance (12-14.5v)
Linear current-regulator ICs are available, but they are as inefficient as resistors when you get down to it. I recently bought a batch of LM 2575T adjustable switching step-down regulators to
experiment further with efficient current regulation for LEDs. I'm sure we all agree that efficient active regulation is the best possible solution. But there's not that much available commercially for the 12v application, and they have drawbacks like complexity, fragility, RF
noise and cost. There's still a place for the lowly resistor in LED lighting from 12v.