Led 3-3.6V what to calculate with? 3.3?

It says it all in the title, if cree leds datasheet says 3-3.6V what do I calculate if want to design serial leds, 3.3V? Or they use whats available? It can be used with 9V supply without resistor or with 12V supply with some resistors?

Modern cree LEDs are constant current, voltage will vary based on thermal resistance.
While many lights have been built with direct drive (battery + resistor) many other use CC drivers.
taskled.com is a good place to start.

Smitiz said:

It says it all in the title, if cree leds datasheet says 3-3.6V what do I calculate if want to design serial leds, 3.3V? Or they use whats available? It can be used with 9V supply without resistor or with 12V supply with some resistors?

Click to expand...

All LEDs are "current controlled" devices. Their voltage is variable based on heat, age, individual differences, etc. If the Cree datasheet says 3-3.6, then each one may be 3, or 3.6, or anywhere in between, those are the outer limits (any produced outside that range are discarded or not sold as the standard product).
So bottom line, you simply can't do what you are talking about, without at a minimum testing, and realistically, you need a constant-current supply. (Many Cree LEDs are at, or below, the lower limit of their specs).

Having said all that, if you run 3x LEDs in series (nominally 9+ volts) and you are talking about a 9V alkaline battery, you have nothing at all to fear, a 9V battery cannot put enough power out to destroy 3 power LEDs, they can't generate the needed current, even if the shelf voltage is above 9V, it'll fall instanly below the safe LED levels when connected.

If you are talking a 9V, or 12V, wall-power transformer, you'll likely blow the LEDs, possibly even with the 9V, as the instantaneous connected current of most 9V transformers is WAY higher at initial connection. An LED can blow in nanoseconds, if the 9V wall-transformer is floating at 16V "open circuit", and connected, it'll dump way too much current into the LEDs. They MIGHT survive, they might not.

Bottom line, you really need some form of constant current regulator, or need to know for sure exactly what your power source is going to do, test the LEDs in series at their operational voltage to match desired current, then at the minimum pick a resistance that matches all that.

Hi,

I have similar question. I want to build an LED light for my Car/Mobike. And to keep things simple, I have thought of this.

Putting 4 LED in series with rating of 3.2 - 3.4V = 12.8 to 13.6 V

a) At stand still Car battery = 12.x Volt. So my LED are safe at engine off.

b) With Engine start = between 12.x upto 13.6 Volt. So, LED are safe even when terminal voltage goes as high as 13.6V.

Problem is that with engine ON, the alternator feeds a lot of current to charge the battery.

Q: Should I be worried about current here? Battery can deliver as much as 20 Amp current but why would it if the LEDs only want less then 1 Amp.

I am totally new here, so need help.

^^ No answers on above ??

Voltage and current in an LED are not unrelated. In fact, as you increase the current, the voltage increases. Likewise if you increase the voltage, the current increases. They go hand-in-hand. The exact relationship depends on the LEDs you choose, the bin you get, and the temperature you run them at. So you need to think of the forward voltage of the LED as a range, not a fixed point. You have some control over the operating voltage, and/or current of the LED, but not complete control over both.

The current drawn by the LEDs is determined by a number of factors. The current available from the source may or may not be one of those factors, and in a car as in many other applications, the current available is well over that needed to fry the LEDs. So this is not a factor limiting the current in the LEDs. Instead the current will be limited by the vehicle voltage, the LED voltage (which is dependent on current and temperature), and the resistance of the wires and connections between the alternator and the LEDs.

This may well protect you in normal circumstances. In fact, you may actually get less current than you expect, unless you are careful about your wiring. I would recommend that a a minimum, you measure the actual voltage in you vehicle when running with a well-charged battery, and measure the forward voltage of your LEDs at operating current and temperature. Compare the two and see if you still think you are okay.

But I think you should be worried, just not for the reason you think. When you do things like turn off the headlights, the voltage can surge to well over 14V. In fact, this so-called 'load dump' can go over 40V! I wouldn't guess that LEDs would last very long if directly connected to the battery voltage in your vehicle, but I really don't have any experience in the matter. I know that minor load dumps happen all the time (like when turning off your headlights), but I really can't say how likely it is that you will fry your LEDs. But I do know that automotive electronics designers take transients, load dump, and 24V jump-start very seriously.

to follow up on the subject of the voltage spikes that are seen in automotive electrical systems, take a look at the this application note from Texas Instruments:
http://www.ti.com/lit/an/snva190b/snva190b.pdf

As a guy who is involved with electronics on slightly larger vehicles, I can tell you that this is just the tip of the iceberg.. or lightning storm? There are lots of types of transients, including negative voltage spikes of hundreds of volts.

Regarding the OP's question... calculate the resistor value using the minimum value for Vf. The current will be lower for all other values of Vf, which will be safe for the LED. If you use a resistor to control the current, then the current won't be regulated or really controlled. It will just be limited. If you want a controlled LED current, then some form of electronics will be needed.

Smitiz said:

It says it all in the title, if cree leds datasheet says 3-3.6V what do I calculate if want to design serial leds, 3.3V? Or they use whats available? It can be used with 9V supply without resistor or with 12V supply with some resistors?

Click to expand...

Probably your best bet is to use a buck/boost (CUK, sepic etc), they're a bit more complex than buck or boost converters but you can take a supply voltage less, the same or more than the load LED requires, a simpler solution is to use a boost converter to up your 3.3V to a more useful voltage (like 12V), then you can use a constant current output LED driver.

Thank you so much DIWdiver , Steve and ianfield.

Very helpful information out there.

In the meantime I have read few more threads on same topic and I TOTALLY agree on the voltage spike part and agree that controlling Automotive voltage, current is a bit too complicated for me.

However I am going to go ahead with my make based on following assumptions and take a chance,

1) Vehicle is Motorbike with 7AH battery and a not so powerful alternator. A Car has much bigger problems.
2) I only switch on my circuit after cranking.
3) On the same vehicle, I am already running 3 LED in series. These 1W LED are much over-driven at 4.5V, they do heat up, but have lasted 1 month and still strong.

This time, I will 4 in Series with 2.2Ohm Resistor. And 4 of such sets will go in parallel making it a 16 LED array. (4X4).


I am happy under-driving them 99% time and ready to take spikes once in a while. Ready to take my chance.

EDIT: One last suggestion I need is: Am I better off using a 1.5W LM317 rather then the resistor. Remember I am using 4 LED in series with current rating 3.2-3.6V.


Reason 4) Cost: total cost of my circuit will be less then 6$. It will be damn easy to replace if needed.


I am not finding enough time to start soldering them, once I do, I will update the results here. I am totally loving the knowledge I get on the forum.

ianfield said:

Probably your best bet is to use a buck/boost (CUK, sepic etc), they're a bit more complex than buck or boost converters but you can take a supply voltage less, the same or more than the load LED requires, a simpler solution is to use a boost converter to up your 3.3V to a more useful voltage (like 12V), then you can use a constant current output LED driver.

Click to expand...

Edit again. My multimeter showed my car and Bike Battery terminal voltage as 13.5V (Turned off) and as much as 15.5V (When turned on).

jeetcp said:

....

EDIT: One last suggestion I need is: Am I better off using a 1.5W LM317 rather then the resistor. Remember I am using 4 LED in series with current rating 3.2-3.6V.


Reason 4) Cost: total cost of my circuit will be less then 6$. It will be damn easy to replace if needed.


I am not finding enough time to start soldering them, once I do, I will update the results here. I am totally loving the knowledge I get on the forum.




Edit again. My multimeter showed my car and Bike Battery terminal voltage as 13.5V (Turned off) and as much as 15.5V (When turned on).

Click to expand...

LM317 vs a resistor? It depends on how you would be using the LM317. Can you sketch out the schematic for the two options? If I had to guess, I'd guess that you want to use the LM317 as a current source. In that configuration, the LM317 requires a certain amount of voltage drop across it.. I think it's around 2v for the dropout voltage (on page 6 of the datasheet that I'm looking at), but I'm not sure if you have to add the 1.2v reference voltage to that or not. If your voltage budget allows for 3.2v across the LM317, then you can be confident that it'll do a better job of current regulation than a resistor. The downside is that it is more vulnerable to nasty voltage spikes than a resistor is.

The battery is getting up to 15.5v when charging? Boy.. that seems pretty high. I was thinking that 14.4v was a pretty agressive charging voltage for a 12v lead-acid battery.

The LM317 can be set up as a current regulator, with just one resistor. That would be ideal for the LEDs. You will need 2-2.5V for the regulator, just be aware of that. You can get one that will handle 60V spikes at the input. Put a 40-50V unipolar transient suppressor in front of that and it would be pretty rugged.

DIWdiver said:

The LM317 can be set up as a current regulator, with just one resistor. That would be ideal for the LEDs. You will need 2-2.5V for the regulator, just be aware of that. You can get one that will handle 60V spikes at the input. Put a 40-50V unipolar transient suppressor in front of that and it would be pretty rugged.

Click to expand...

The sensing resistor for a LM317 current regulator has to drop 1.25V, quite a power loss in a power LED driver - with a 2 transistor current limiter, the sense resistor only has to drop 0.7V.

I like the two transistor current regulator myself (it's a classic and can be found in lots of places), but I'm not sure the OP has the skills to implement it.

I really dont. And I dont even know how to add LM317, but I will read a little bit around, I am sure it is very common thing. Let me get back after this guys. I am so liking it.

Steve K said:

I like the two transistor current regulator myself (it's a classic and can be found in lots of places), but I'm not sure the OP has the skills to implement it.

Click to expand...

Steve K said:

I like the two transistor current regulator myself (it's a classic and can be found in lots of places), but I'm not sure the OP has the skills to implement it.

Click to expand...

Its so simple I can describe it adequately without a schematic: Take an emitter follower and put a sensing resistor in the emitter lead,, add a small signal type transistor with its emitter to the output end of the emitter resistor, its base to the junction of resistor & power transistor and its collector to the base of the power transistor. Then all that's required is to calculate the value of resistor to drop about 0.65 - 0.7V at the desired current.

but that still assumes that the OP knows what a transistor is, how to figure out which ones to buy, etc.

Since the subject comes up periodically, I did post a quick schematic to my flickr account a while back, with some very brief instructions on calculating some resistance....
http://www.flickr.com/photos/110121...BLk-bgRFkD-9HnZ3g-b5Jqai-b5JidX-9HqRvE-beEtT8

Steve K said:

but that still assumes that the OP knows what a transistor is, how to figure out which ones to buy, etc.

Since the subject comes up periodically, I did post a quick schematic to my flickr account a while back, with some very brief instructions on calculating some resistance....
http://www.flickr.com/photos/110121...BLk-bgRFkD-9HnZ3g-b5Jqai-b5JidX-9HqRvE-beEtT8

Click to expand...

You want me to go and build it for them!

that would be a generous thing to do.

I think this gets to a basic problem on any sort of techie web forum... people are looking for solutions to problems, but really lack the fundamental knowledge to do more than plug in a standardized product. It is conceivable that a person might put together a nice video describing the circuit, providing digi-key part numbers, maybe using all leaded parts so a simple perf-board would be needed to build it up, etc., but that's way more work than I'm willing to put into something like this.

The simple LM317 current regulator circuit is simple enough that the builder wouldn't need much more than the ability to solder a leaded resistor to the LM317 leads and attach wires. ..and possibly a heatsink.

Honestly, there are more than a few threads where I think I know of a technical solution, but it involves building a circuit that appears to be far beyond the OP's capabilties. I usually just watch from a distance.