# LED Dimming in Parallel

Discussion in 'The Projects Forum' started by pfhorge, Feb 23, 2007.

1. ### pfhorge Thread Starter New Member

Feb 23, 2007
4
0
I'm trying to use an RGB LED to make a dimmable lamp for myself, but I'm pretty confused. The LED I'm using is a Z-Power P1 (spec document here). I'm putting a pot in series with each LED and resistor, so there are 3 sets of pot--LED--resistor in parallel. I've read plenty of sites in an attempt to get my head around this, but I'm confused for several reasons.

First, this LED package shares a single anode between each LED element in the package. All of the sites I've read show that the resistor that goes in series with each LED should go on the anode side of each LED, which I obviously can't do here. If I put the resistor on the cathode, won't that drop the voltage to my LED? In order to get ~.35A to each LED branch of my circuit, I think I need about 13 Ohms of resistance, which would drop all of my voltage before the LED.

Second, I did a project to make a mood lamp (link), which is why I have one of these Z-Power LEDs lying around. I'm looking at the resistor values that guy chose in his project, and they don't make sense to me. With his 3 resistors, the total resistance of his circuit should be something like 1/((1/5)+(1/10)+(1/15)) = 2.7 ohms. With a 5 volt power supply, he should be drawing 1.85 amps, but he only calls for a 1 amp supply. Does his circuit only work because the power supply is hitting its limit? I understand circuit math better now than I did when I did the project, so I'm confused about why something that I know works doesn't seem to meet up with my understanding of the math.

Also, I have a general LED question. LEDs are current controlled devices, so what happens if there's not enough voltage at a given current? Does the device just get dimmer, or does the LED not get enough voltage to conduct, or what? I'm wondering this because it's not entirely clear to me how a potentiometer and an LED interact, since one controls resistance/voltage, and the other is controlled by current.

I've read over this post several times, and I keep halfway convincing myself that I know an answer to a given point, but I think I'd rather have some real advice instead of smoking a \$20 LED. Any help would be appreciated!

2. ### beenthere Retired Moderator

Apr 20, 2004
15,815
283
This is where an understanding of both LED's and the way currents divide in parallel branches would be very useful.

With a common anode, it is possible to place a current limiting resistor common to all three LED's. This might not suit your proposed application, as a mood light needs to let you control each LED separately.

So you will need a fixed resistor in each parallel branch that will limit current to a safe maximum, as determined by the spec sheet. Further variance in brightness will come from a potentiometer(digital or manual) in each branch, set up as a variable resistor.

The values of the resistances are determined by the formula I = E/R. The E you work with is the voltage left after you subtract the individual LED's voltage drop from the applied voltage. Current in the parallel branches, and therefore through each LED, may be controlled independantly of the other two branches.

3. ### pfhorge Thread Starter New Member

Feb 23, 2007
4
0
An electrical engineer friend of mine came by over the weekend, and we got things sorted out. I was confused mainly by the idea that an LED could have enough current, but not enough voltage, and thus be on but dim.

Also, my current version of the circuit suffers from needing some fairly strong pots to dim the LEDs enough. The high resistance pots mean that the LEDs increase in brightness modestly for about 75% of the pot's rotation, then get blindingly bright in the last 25%, so fine adjustments can be tricky.

I think for the next version I'm going to go back to the microprocessor approach. I'd like to try out an AVR (didn't much like PIC assembly), and I think this would be a good place to try it. I understand the circuit well enough to feel comfortable adding a couple of transistors and developing a little code. PWM just seems like the way to go for the level of control I'd like.

Thanks for the help!

4. ### beenthere Retired Moderator

Apr 20, 2004
15,815
283
Or use log response pots instead of linear.

5. ### Ron H AAC Fanatic!

Apr 14, 2005
7,050
657
Actually, that can't happen. The voltage and current are inexorably tied to each other. If it has enough current, it has enough voltage,

There was some mention of using one resistor to control the brightness of all three by putting it in the common anode circuit. If you tied all the cathodes together, this will definitely not work. It is not even a good idea for 3 LEDs of the same type, and certainly not for RGB. The current will be hogged by the one with the lowest voltage drop, leaving the others dim or even dark.
PWM is definitely the way to go, but keep in mind that you still need a resistor in series with each LED.

6. ### pfhorge Thread Starter New Member

Feb 23, 2007
4
0
I think I am confused by a graph in the spec document of Forward Current vs. Forward Voltage. Does the forward voltage vary with the forward current? The graph appears to say one thing, but the tables of data seem a bit different. Page 10 of that document shows that at 350mA for green, the forward voltage should be a little less than 3V. However, on page 7 it says that at 350mA the forward voltage should be 3.5. Is this just an error?

I was looking at the graph and thinking that it was possible to be at a given current and not have enough voltage, but I suppose if that were the case you wouldn't get the diode to conduct, right?

7. ### Ron H AAC Fanatic!

Apr 14, 2005
7,050
657
You're right, their datasheet is ambiguous. However, if you were to take an LED (or diode, or resistor) and sweep the current from zero to some maximum, you will get a unique voltage for each value of current, such as the curve on p.10. For resistors, this is known as Ohm's Law.
Of course, curves for different devices will not all look like the one on p.10, but for any device, for a given current, you will get the voltage that corresponds to that current. You can't change current without the voltage changing, and vice-versa.