/diy/ - Do It Yourself

[Anonymous No.467145]

I bought a few white LED's 3V 350mA thinking I would build a 16 LED panel for my garage using 4 in a series and supplying it with 12V DC.

It seems that it is not as simple as I thought and that I need a driver for it as shown here http://www.instructables.com/id/Super-simple-high-power-LED-driver/#step1

Having never studied any of this I am very confused. I was suggested using one of those many LED calculators available online but its all gibberish to me. These are the calculators I was pointed to
http://led.linear1.org/1led.wiz
http://ledcalc.com/
http://www.ledcalc.net/?betriebsspannung=12&led_anzahl=6&led_spannung=3.5&led_strom=350

The original link said I needed an 9ohm 1W resistor for 1 LED. One of these calculators tell me I need a 33ohm 4W resistor for 1 LED.

Can someone please simplify this and tell me which resistor I need for 4 3V LED's in a series powered by 12V DC. What if I decide to use 6 2V LED's in series, do I need a different resistor in that case?

[Anonymous No.467157]

Get a 12V 350mA wall wart. There you go.

Make sure those LEDs are on heat sinks

[Anonymous No.467175]

>>467145
http://www.overclock.net/t/47351/how-to-use-resistors-and-leds-in-my-rig-a-guide-for-the-first-timer

[Anonymous No.467197]

>>467145
The voltage rating of an LED does NOT mean you can just apply that much voltage and expect it to work - it means that when set up properly that's APPROXIMATELY how many volts will be dropped across it. You can then choose a resistor to get approximately the right current through the LED. For example say you have a 2V 100mA LED, and a 45V power supply.

You assume there's 2V across the LED. that means there will be 45 - 2 = 43V across the resistor. Now use ohm's law: R = V/I, we want I = 100mA ( = 0.1 A), so R = 43/0.1 = 430 Ohms. Now the power dissipated across this will be calculated by P = I * V. So 43 * 0.1 = 4.3W, So you need at least a 5W resistor, or to be safe, use an 8W or 10W.

If you connect lots of LEDs in series like you were suggesting, the percentage error of voltage across the resistor becomes too much to be able to control the current properly without knowing the characteristics of the LED, accuracy of the resistor, etc. So it's better to connect LEDs in groups that still allow some room for approximation in the calculations. So in your case, connect the LEDs in paralled or into pairs and then into parallel, and then give each pair its own resistor, and then do the calculations.

tl;dr LEDs are non-linear devices and require resistors to regulate the current to them.

CAPTCHA: engineer tyaIBas

[Anonymous No.467207]

>>467157
I cant simply use 16 wall warts.

>>467175
Thanks that was helpful.

>>467197
Okay that makes some sense. Gonna try doing that tomorrow.

Most of the articles I read online say that LED's need a Constant Current driver. What is that about? Is a lead acid battery considered a Constant Current?

[Anonymous No.467213]

>>467207
Most of the articles I read online say that LED's need a Constant Current driver. What is that about? Is a lead acid battery considered a Constant Current?

No. Constant Current Source is probably what they mean. That's usualy a transistor setup so that it can vary it's voltage output to keep the current across the LED constant. It's more efficient than just dumping current into a resistor, but more complicated. C'est la vie.

[Anonymous No.467219]

>>467145
Here's the thing: LEDs are driven by *current*, not *voltage*; you really need a current-regulated voltage source for them, not a voltage-regulated current source. Why? Because as they heat up (and there is indeed self-heating from power dissipation) the voltage drop across the junction will change. You can use an LM317 configured for current regulation.

[Anonymous No.467223]

>>467213
Actually, if you are using a linear current regulator, it will be the same efficiency as a resistor, seeing as a regulator acts like a resistor that will change on its own to give you a constant current. A fixed resistor is less accurate, and for most practical purposes, a better option.

[Anonymous No.467230]

>>467219
I was gonna go for it but it wastes 3V from the power source. Not very efficient.

Is there no other way than the LM317?

>>467223
If I use just a resistor without a variable voltage regulator will it work? Or the LED's will blow up because of the heat?

[Anonymous No.467240]

>>467230
It depends: If you have 4 leds hooked up in series with a very small resistor or no resistor at all like the OP, then as the LEDs heat up they will draw more current, get hotter, etc. and then blow up.

So the setup must be chosen such that there is enough voltage dropped across the resistor that it doesn't 'care' about what the LED is doing, ie if the LED heats up, it's so called 'impedance' will go down, but the current increase will be negligible because the resistor is the main source of regulation.

[Anonymous No.467261]

>>467213
You really don't want to use a simple resistor circuit with high powered LEDs. The resistors can get hot enough to melt solder as well as overheating nearby components such as the LEDs themselves. Constant current drivers are the way to go.

Also is your 12v source a set of batteries? Those drop in voltage over time. LEDs aren't like lightbulbs that gradually dim with a lower voltage, they will not run at all if the voltage is slightly below specification.

[Anonymous No.467264]

>>467145
i hope she gets fucking gored

[Anonymous No.467287]

>>467264
I'd fucking gore her, if you know what i mean

[Anonymous No.467289]

>>467287
RIP The usefulness of this thread...

/thread

[Anonymous No.467334]

http://www.instructables.com/id/Power-LED-s---simplest-light-with-constant-current/#intro
Then just aim for 5.6A, or bundle the LEDs in x4s to make it a whole fuckton easier on yourself.

[Anonymous No.467545]

>>467240
>>467261

From my limited understanding of Electronics I know that voltage is pushed while current is pulled. Is this not the case with LED's? Why will they be pulling more current than they need?

Yes I was thinking of using my spare lead acid battery and hook a solar panel for my first solar experiment then add to after I learn some more about it.

>>467334
Lets see how it goes, I have ordered several resistors and voltage regulators. Fingers crossed for a next day delivery

[Anonymous No.467548]

>>467545
>Why will they be pulling more current than they need?
They're diodes. As long as the voltage is the right way around, they will have near-zero resistance (short). This will quickly cause them to heat up and let the smoke out if you don't limit it with an external resistor.

[Anonymous No.467555]

>>467545
>Why will they be pulling more current than they need?
Incandescent lights are like resistors. As more current goes through them the voltage drop across the light increases. Thus if you put a constant voltage across it the current going through is limited by the resistance.
LEDs are not like resistors. A very small change in voltage will lead to a very large change in current. For example, maybe if the voltage is 2.0374 the led will let a safe 10mA through, but if the voltage is 2.0375 it will let 500mA through, which will quickly burn it out. These voltage characteristics vary with temperature and are not reliable enough to use. Instead, you can put a resistor in series. The current passing through the resistor and the LED is the same, so using ohm's law you'll find that when the current increases the voltage across the resistor increases, and by KVL the voltage across the LED will go down. In this way the resistor will keep the current from spiraling out of control.

[Anonymous No.467562]

>>467223
Except you don't have the voltage drop across a resistor, surely? You lose your 0.7V and some current. Unless your saying Rb is going to be like 1K. Which ain't likely.

There's a good article on IC based LED driving here. Inductor based drivers seem to be the most efficient because of the capactive loading of LED's, but they don't operate so well at voltages above 20V.

http://www.maximintegrated.com/app-notes/index.mvp/id/1804

[Anonymous No.467579]

Anyone suggesting to run LED in parallel forgets one key thing.

The path of least resistance.

If you want equal current to go across the LEDs in a parallel config, you will need to add in some additional circuitry.

[Anonymous No.467602]

>>467230
>If I use just a resistor without a variable voltage regulator will it work? Or the LED's will blow up because of the heat?
That's one way of doing it, but using a current regulator is better.

[Anonymous No.467603]

>>467230
>Is there no other way than the LM317?
Sure but that's the most accessible solution for the typical level of expertise on /diy/. You want to go find a PWM driver or switching power supply circuit that's current regulated? Go for it.

[Julez No.467651]

http://www.ebay.de/itm/380464617528?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1438.l2649

This one is what you need. Input can be connected to 12VDC, output goes up to 60VDC, current can be regulated from 0-2A, so 350mA are in that range.

[Anonymous No.467837]

>>467651
That is rated at 60W, does this mean I could theoratically use 60 of the 1W LED's?

>>467603
Pretty sure I am gonna blow all these up. Already adding stuff in my basket to order next week. Hopefully it will go better the second time.

>>467602
Will experiment with it. Thanks

>>467548
>>467555
That sucks. Would have been awesome if it had been simpler, or if I understood some of it.

[Anonymous No.467850]

>>467837
TL;DR

Diodes are silly, don't trust them. They'll require a certain voltage and current. Let a regulator decide that for you and don't worry about their complex inner lives. Fully understanding it is great, but outside of the scope of this project other than to justify while the simpler methods have numerous dodgy fault conditions (situations in which something bad happening makes more bad things happen/brings device to a halt).

[Anonymous No.468304]

>>467850
Can I use this instead of an LED driver?

http://english.bincolor.com/Product32/BC-321-10A.html

[Julez No.469018]

Hi!

U can't use the BC-321-10A driver you linked. Take the Ebay link from me instead, because it will work.
Diodes are non-linear devices. If you are near the operating point at 3V, then a little more or less voltage will make a huge difference. Therefore, you need a driver, which regulates the current in the range you want, and is able to provide the voltage the diodes need.
White LEDs require around 3V to work. The current can be, depending on the LED and cooling, between 5mA (tiny signal diode) and 5A (Cree XM-L2 with copper heatsink andsuper cooling).

Your LEDs with 350mA rated current and 3V are around 1W each.
(0.35A*3V=1.05W)
60 of them in series, however, would require 180V.
(3V*60=180V)
The ebay driver can provide only 60V, so you can operate 30 LEDs maximum.
(60V/3V=20)

16<20, so your 16 LEDs will work with that driver.

Just buy this driver, and I can guarantee that it will work.

Always remember #1 rule of engineering:
"If you don't have a clue, do it exactly the way others do. Only when you have succeeded here, you may try a different way."

And I can say, I have a clue.
Here is one of my LED lights.
Old version:
http://imgur.com/a/cfG9T#0

New version:
http://imgur.com/a/voKaM#1

[Julez No.469021]

I meant:

The ebay driver can provide only 60V, so you can operate --20-- LEDs maximum.
(60V/3V=20)

[Anonymous No.469058]

i have a bunch of older solar lights from a yard. the batteries (AA) inside are screwed but the LEDs and the small panels still work. is there an easy way to attach all of these panels together to made one bigger panel? or would i need more stuff?

[Anonymous No.469065]

>>469058

You can run the panels in parallel to increase the amperage, be careful running them in series as you'll encounter heavy efficiency drops with every shitty solder joint and panel you add.

>>469018
>Diodes are non-linear devices.
Regular diodes, yes. LEDs are a slightly different story, depending upon manufacturing method.

[Anonymous No.469094]

>>469018
Alright thanks. Will get that driver you linked.

Your setup seems awsome. Can you provide the specifications on those LED's? Wattage and luminosity

[Anonymous No.469100]

>>467145
God damn OP that picture is hot as fuck

[Anonymous No.469103]

>>469100
if only it wasn't upscaled and at the wrong aspect ratio

[Anonymous No.469108]

>>469103
There's loads of versions if you reverse image search it, but I still don't know which the original is. Probably not this one, too fuzzy. Can't get sauce either because a load of people used this picture for farming simulator or something.

[Julez No.469113]

Hi!

These are the LEDs I use:

http://www.wayjun.com/index.php?main_page=product_info&products_id=856

I run them at 680mA. Have forgotten the exact specs, but I get about the equivalent on 300W incandescent light, with around 40W electrical input.

I use rather cold light (6500K) for work. Keeps you from getting tired. Choose warm white for living room or when you like a cosy atmosphere.

However, I would recomment these clusters now:

http://dx.com/p/60w-6000k-4800lm-20-led-white-light-bulb-silver-dc-30-36v-155743

Epistar is good tier, but Cree is god tier.
And, underdriving is the key to longlivety and success.

[Julez No.469115]

I forgot, underdriving means operating the LEDs below their max specs.
Like 680mA instead of 1.7A.

[Anonymous No.469118]

>>469115
How much do their brightness suffer because you underdrive them?

Have you ever bought something from dx.com? Do they have premium shipping? A friend apparently bought something with free shipping which was lost in the mail and their support was terrible.

[Julez No.469119]

And, as a last advice:

Hang out at candlepowerforums.com.
This is the place in the internet where the real pro's gather discussing LEDs and stuff.

[Julez No.469120]

The brightness suffers a bit less than linear, like 60% brightness at 50% current. But brightness suffers also when the LEDs get hot.
So, by underdriving them, you end up losing less than you might expect, as efficiency increases with smaller current. Getting the heat away is one of the most difficult tasks when building a LED setup. My lampshade gets too hot to touch after a while, but as I underdrive the LEDs, they can cope with the heat. Therefore I recommended the driver with the adjustable current. You can still reduce it if the LEDs get too hot for your liking.

I order from DX a lot. One order was messed up in the beginning, but all others were flawless. Takes between 2 weeks and a month to arrive in Germany

[Anonymous No.469126]

>>469120
Do you choose their free shipping?

One last question. Can you have a look at this page http://www.ledlights.uk.com/accessories-page-5.html

I am confused at what those LED power supplies. What are they? Some kind of driver?

[Julez No.469168]

Hi!

I always had free shipping. In general, every item I paid for on the interwebz finally arrived. Call me lucky :-)

Your link: These are not drivers. Drivers have a regulated current, and the voltage the LEDs need is automatically provided (within the range of the specs).

The power supplies you link have a fixed voltage, and will supply the current the connected load needs. One cannot hook LEDs directly to them. There are so called "LED strips" which are designed to run on 12V. Their current is kinda regulated with a small resistor in series to 3 LEDs. For every 3 LEDs on such a strip, there is one resistor. So you can put these strips directly on any 12V DC source, be it a car or your linked power supplies.

LED strips are awesome to put under your kitchen cabinets.
My advice: Buy a high density strip (120 LEDs per meter) of cold white and neutral/warm white each.
Put them in parallel under your cabinets, and connect to a cheap 12V/5A laptop power supply for like $10.
The light temperature and brightness will amaze you! And the installation is really easy.

[Julez No.469176]

http://www.youtube.com/watch?v=UqRHw_0ueaQ&feature=player_embedded

That's what I mean.

[Anonymous No.469734]

[Anonymous No.469741]

>>467145
>Having never studied any of this I am very confused.
Perfect opportunity to pick up a book and start.

[Anonymous No.469790]

>>469741
Too boring. Learning by trial and error is much more fun. Blew up a few bulbs, short circuited a battery making it useless and burned a finger while soldering. I think now I know exactly what to do.

>>469176
Thanks man, you were very helpful. I am gonna have an awesome LED panel in the next few days which works on solar.