Need some EE students help (or electricians, or tinkerers)

I'm trying to figure out a way to create my own low power LED lighting system for a room. I apologize in advance for the amount of stupidity that will follow. I haven't used my electrical knowledge in a decade or so. And even back then it sucked, but I digress.

I've already (in general) got the plan laid out. It's the final steps that I'm having trouble figuring out. I plan on using 100ea 3mm Bright White LED's. The specs on the LED's are 20mA, 15 lumens, 3.3v.

Here's my plan:

Code:

---

+________+_______+_______+_______+_______+__
PWR        LED        LED        LED        LED        LED
-________-_______-_______-_______-_______-__

---

Basically, using some of that "flat" speaker cable as the positive and negative leads, and simply soldering on the led's every few cm's or so.

So that is parallel, correct? I found this page which seems to indicate that I am wiring them in "parallel", so I'm going to assume that is correct.

Then I have a 3.3v circuit that will need 2A to drive it, or does the current add up differently?

What is an efficient 3.3v source? Or would it be better to sting a few of these in series to bring it up to 12v?

You can use a higher voltage source but you'll have to a current limiting resistor with each one.

If you're using a 12V source, the LED will drop 3.3V and your left with 8.7V. Limiting that to 20mA requires a 435 ohm resistor. A 470 (common size) will give you 18.5 mA which should be fine.

You could group (3) LED's in series which results in a drop of 9.9V. Having 2.1V left, a 105 ohm resistor. You can get a 110 ohm for 19.1mA. 1/4W resistors are fine. You're dissipating about 40mW on the resistor with the three LED's.

P.S. I don't like parallel arrangements with a single resistor as the LED's are not all made equally.

Wow, go Adisharr and you're crazy electrical knowledge. :D

I don't really enjoy the idea of soldering in a resistor on each LED.

So in theory, if I grouped them in groups of 4, I could roughly get away with no resistor? (4x3.3=13.2v) They wouldn't run as brightly, but they should run, or am I undervolting them too much at that point?

Another question, would this setup be dimmable? Or do LED's simply shut off after reaching a specific low voltage?

hmmm

oki.. in general , it is not WISE to attach any Lights or Leds in serial .. even in groups of serials attached parallel

why?

one ..well if one dies the whole group will turn off on that branch (bad for maintainance and determining which is faulty and looks bad)


two.. u will need a very very high volt source and low current which is dangerous (high volt)

just use parallel and use the Voltage needed in ur case the total current is the sum (parallel)

usually we use WATT to calculate the powersupply needed .. WATT is independent u will need SUM(each part watts in use) easy and simple.

You didnt say why u need 12V source??? I dont see any reason for that just use 3.3V power source why complicate things? why loosing power into heat???

Edit: dont get exact power need .. get like 25% more .. in your case like 3.3V 2.5Ampere power adapter.

Quote:

---

Originally Posted by fist

hmmm

oki.. in general , it is not WISE to attach any Lights or Leds in serial .. even in groups of serials attached parallel

why?

one ..well if one dies the whole group will turn off on that branch (bad for maintainance and determining which is faulty and looks bad)


two.. u will need a very very high volt source and low current which is dangerous (high volt)

just use parallel and use the Voltage needed in ur case the total current is the sum (parallel)

usually we use WATT to calculate the powersupply needed .. WATT is independent u will need SUM(each part watts in use) easy and simple.

You didnt say why u need 12V source??? I dont see any reason for that just use 3.3V power source why complicate things? why loosing power into heat???

Edit: dont get exact power need .. get like 25% more .. in your case like 3.3V 2.5Ampere power adapter.

---

I'm not sure you're all that familiar with the behavior of semiconductors. All LED's are going to be slightly different - even the same type in the same batch.

A small voltage variance can cause a large change in brightness. Connecting a 3.3V LED across a 3.3V source can work but if your supply voltages happens to change by +0.1V you could be looking at a current flow of DOUBLE what you had at 3.3V.

This problem is eliminated with limiting resistors.

By 'very very high volt source' I'm guessing 12V is high voltage in your country? As far as why I talked about a 12V source - your should actually read what I wrote. That's for three LED's in series.

As for one damaged LED inhibiting the rest, you're correct. I've had LED's running for years w/ no problems so I don't think that's much of an issue. I'm trying to save time on some initial wiring. Using three in series results in 33 resistors vs. 100 for having one on each LED.

For total wattage estimates

*** 12V supply ***

(3) series LED's + 110 ohm resistor = approx. 7.6W total for a (99) LED set-up

(1) series LED + 470 ohm resistor = approx. 22W total for a (100) LED set-up

*** 5V supply ***

(3) series LED's - not applicable

(1) series LED + 91 ohm resistor = approx. 9.2W total for a (100) LED set-up

As you can see the 12V set-up uses less power and is easier to wire.

Quote:

---

Originally Posted by James

Wow, go Adisharr and you're crazy electrical knowledge. :D

I don't really enjoy the idea of soldering in a resistor on each LED.

So in theory, if I grouped them in groups of 4, I could roughly get away with no resistor? (4x3.3=13.2v) They wouldn't run as brightly, but they should run, or am I undervolting them too much at that point?

Another question, would this setup be dimmable? Or do LED's simply shut off after reaching a specific low voltage?

---

I've used some LED's in the past but I'm sure not an EE ;)

The (4) series set-up would reduce the current through each LED to around 5mA which would be very dim.

You could dim the whole set-up if you put a potentiometer in series with the power supply. I don't have any brightness output vs. current graphs though. I believe it's going to be non-linear but usable.

Interesting.

Obviously there are some things I"m missing.

So would there be no way to wire them all up in parallel without using resistors at all? Or is that not possible?

I know that there will be variances from LED to LED, but my question is will that truly cause that many issues?

Would it be worth testing a 10 LED string with this? The price of the LED's is (relatively) trivial, so it would be worth it to me to try it, even if I fry some. Especially since I got my job. :D

Quote:

---

Originally Posted by Adisharr

I'm not sure you're all that familiar with the behavior of semiconductors. All LED's are going to be slightly different - even the same type in the same batch.

A small voltage variance can cause a large change in brightness. Connecting a 3.3V LED across a 3.3V source can work but if your supply voltages happens to change by +0.1V you could be looking at a current flow of DOUBLE what you had at 3.3V.

This problem is eliminated with limiting resistors.

By 'very very high volt source' I'm guessing 12V is high voltage in your country? As far as why I talked about a 12V source - your should actually read what I wrote. That's for three LED's in series.

As for one damaged LED inhibiting the rest, you're correct. I've had LED's running for years w/ no problems so I don't think that's much of an issue. I'm trying to save time on some initial wiring. Using three in series results in 33 resistors vs. 100 for having one on each LED.

For total wattage estimates

*** 12V supply ***

(3) series LED's + 110 ohm resistor = approx. 7.6W total for a (99) LED set-up

(1) series LED + 470 ohm resistor = approx. 22W total for a (100) LED set-up

*** 5V supply ***

(3) series LED's - not applicable

(1) series LED + 91 ohm resistor = approx. 9.2W total for a (100) LED set-up

As you can see the 12V set-up uses less power and is easier to wire.

---

ah thanks for the info lol...

regarding the serial high volt , I was talking in general not 3 only ... talking in concept not special cases..


again thanks.. ^^

Quote:

---

Originally Posted by James

Interesting.

Obviously there are some things I"m missing.

So would there be no way to wire them all up in parallel without using resistors at all? Or is that not possible?

I know that there will be variances from LED to LED, but my question is will that truly cause that many issues?

Would it be worth testing a 10 LED string with this? The price of the LED's is (relatively) trivial, so it would be worth it to me to try it, even if I fry some. Especially since I got my job. :D

---

You could always try it :) I wouldn't exceed 3.3V though or you'll start seeing a lot of failures.

Yeah I agree parallel arrangement of LED's is never the way to go, you should series them until you are close to the supply voltage and then parallel each chain

Quote:

---

Originally Posted by Adisharr

I've used some LED's in the past but I'm sure not an EE ;)

The (4) series set-up would reduce the current through each LED to around 5mA which would be very dim.

You could dim the whole set-up if you put a potentiometer in series with the power supply. I don't have any brightness output vs. current graphs though. I believe it's going to be non-linear but usable.

---

Wait, I guess I'm missing something. (I really hate to ask you to explain even further than you have.)

How do you calculate the current through the series? (You mention reducing it to 5mA by putting 4 LED's in a chain. What is it for the 3-LED chain?) I know there is some relatively simple piece of information that you obviously are offering me and I'm overlooking.

Finally, does anyone have a good "3.3v" power source they recommend? I know it will probably be a variable one, I just don't even know what to look for.

My first round of testing will probably consist of 10-20 LED's, some resistors, and a lot of "magic blue smoke". ;)

Quote:

---

Originally Posted by James

Wait, I guess I'm missing something. (I really hate to ask you to explain even further than you have.)

How do you calculate the current through the series? (You mention reducing it to 5mA by putting 4 LED's in a chain. What is it for the 3-LED chain?) I know there is some relatively simple piece of information that you obviously are offering me and I'm overlooking.

Finally, does anyone have a good "3.3v" power source they recommend? I know it will probably be a variable one, I just don't even know what to look for.

My first round of testing will probably consist of 10-20 LED's, some resistors, and a lot of "magic blue smoke". ;)

---

Hey James - no worries :)

When you series LED's with a power source and a resistor, you would do the following (3 LED example - assumming 3.3V forward voltage and 12V power):

1. Resistor voltage = Voltage source - ((# of LEDS) x (forward voltage))

12V - (3 x 3.3V) = 2.1V

The 2.1V is what's left which is across the resistor.

If each LED needs 20mA, we need to find the resistance that will allow 20mA to flow with a voltage of 2.1V.

R = V / I

R in ohms, V in volts and I in amps

R = 2.1 / .020

R = 105 ohms

You could use the next closest size resistor or a combination of resistors in series to get that resistance.

You want to make sure that you don't exceed the capacity of your power supplies voltage output. Four LED's will drop 13.2V which is over your supply.

Diodes have interesting behavior - if you look at a graph of current vs voltage on a 3.3V LED, you'll see that the current changes rapidly up or down on small deviations of the supply.

Take a look at figure 1. in this datasheet:

http://focus.ti.com/lit/an/slyt284/slyt284.pdf

The top LED shows a current of about 7mA at 3.0V and about 60mA at 3.6V.

**** Four LED's in series - no resistor ****

If you assume all your LEDs are exactly the same then you'll have an equal voltage drop across each LED.

12V / 4 = 3V

You'll get around 7mA through them which won't light them much.

Ok, thanks.

So it looks like series of 3, with resistors might be the way I have to go.

What I'm trying for is to see if there is a way to create an ultimately low power lighting system that can be "hidden" behind false cornices in a room that allows for the whole ceiling to be a reflective lighting surface.

While individual LEDs aren't that bright, I figure if I string them up around the upper border of a room, it should provide plenty of light, for less power than even a CFL bulb requires and produces significantly less heat.

Quote:

---

Originally Posted by James

Ok, thanks.

So it looks like series of 3, with resistors might be the way I have to go.

What I'm trying for is to see if there is a way to create an ultimately low power lighting system that can be "hidden" behind false cornices in a room that allows for the whole ceiling to be a reflective lighting surface.

While individual LEDs aren't that bright, I figure if I string them up around the upper border of a room, it should provide plenty of light, for less power than even a CFL bulb requires and produces significantly less heat.

---

Not a bad idea James - if you do this I'll require many pictures :D