One resistor in parallel or a resistor for each?

What are the advantages and disadvantages for using one resistor for a group of LED's wired in parallel or having one resistor for each LED.

Does it matter where you put it? (I mean on the positive or negative side).

[ QUOTE ]
flashlightDR said:
What are the advantages and disadvantages for using one resistor for a group of LED's wired in parallel or having one resistor for each LED.

[/ QUOTE ]

Advantage of one resistor: Only one resistor required.
Advantage of multiple resistors: Current to each LED will be more similar than for the case of only one resistor.

So its generally better to use multiple resistors?

[ QUOTE ]
flashlightDR said:
So its generally better to use multiple resistors?

[/ QUOTE ]
IMHO, yes.

That's assuming each led has a different Vf from the others and that you adjust the resistor for that led accordingly. That means taking a curve trace (I-V) of each led to determine the Vf and then setting the resistor current for a given constant input voltage. Kind of a pain in my opinion unless you know that the Vf varies wildly between the leds in question. It's more practical to use a single resistor that is able to handle the power it must drop from the power supply.

[ QUOTE ]
CM said:
That's assuming each led has a different Vf from the others and that you adjust the resistor for that led accordingly. That means taking a curve trace (I-V) of each led to determine the Vf and then setting the resistor current for a given constant input voltage. Kind of a pain in my opinion unless you know that the Vf varies wildly between the leds in question. It's more practical to use a single resistor that is able to handle the power it must drop from the power supply.

[/ QUOTE ]

No, that assumption is not required. If you take a group of LEDs with any distribution of LED Vf characteristics, and ballast them all individually with the same individual resistor value, the resulting current in each will be more similar than if you had used a single resistor to ballast the whole group.

Doug S,

I don't dispute that assertion. The point I was trying to get across is that if you want to have a more accurate (in my example, it is exact) bias, you would use individual resistors tweaked to that particular LED. But in reality, the Vf does not vary so wildly that this is necessary. You could just use one resistor for the sake of economics. Multiple resistors of identical value is more optimal, a single one is more economical while giving similar results. As engineers, we always have to make design compromises and if it can be done cheaper and obtain similar results, then by gosh that's the way to do it. That's my pragmatism speaking out /ubbthreads/images/graemlins/grin.gif

Re: One resistor in parallel or a resistor for eac

[ QUOTE ]
CM said:
Doug S,

I don't dispute that assertion. The point I was trying to get across is that if you want to have a more accurate (in my example, it is exact) bias, you would use individual resistors tweaked to that particular LED. But in reality, the Vf does not vary so wildly that this is necessary. You could just use one resistor for the sake of economics. Multiple resistors of identical value is more optimal, a single one is more economical while giving similar results. As engineers, we always have to make design compromises and if it can be done cheaper and obtain similar results, then by gosh that's the way to do it. That's my pragmatism speaking out /ubbthreads/images/graemlins/grin.gif

[/ QUOTE ]
Multiple identical resistors of the same value will give FAR more consistent results than one resistor and all LEDs in parallel.

For example, suppose you have two Luxeon LEDs, one with a Vf of 3.4v, and one with a Vf of 3.6v. You're powering them from 5V. As an example, you choose resistors so that the LEDs are being driven at their rated current at 3.5v. Resistance for each resistor will be 4.286 ohms.

Now, in reality, one LED will be dropping 3.4v and one will be dropping 3.6v. (Plus/minus a few insignificant hundredths since they will not be at their exact rated current.) So the 3.4v LED will be driven at 1.6/4.286 = 373 mA. The 3.6v LED will be driven at 1.4/4.286 = 326 mA. Variance, but not too much.

Now, if you use a single resistor and parallel the LEDs - Let's just take an average of 3.5v. (Most likely it'll be more towards the lower side.) This means a total of 700 mA.

The problem is that at the node with the + side of the two LEDs and one end of the resistor, the voltage is fixed. Given the nonlinear V/I curve of LEDs, the 3.4v LED will likely be quite overdriven at 3.5v, and the 3.6v LED will be quite underdriven at 3.5v. I would not be surprised to see 600 mA+ going through the 3.4v LED and <100 mA through the 3.6v LED - Even a tenth of a volt can make a LARGE difference in the current through an LED.