3 Watt LED’s In a flood light

Thread Starter

RodneyB

Joined Apr 28, 2012
697
I use the following web site always to work out the array and resistors for LED circuits.

http://ledcalculator.net/

My 3 watt LED’s have the following specifications

Led Voltage Drop 3.6 Volts

LED current 700mA

Number of LED’s 12.

Results from WEBSITE
  • You will need 4 x 1.8 ohm 1 watt resistors.
  • The 1.8 ohm resistor is colour coded: Brown, Gray, Gold, Gold.
  • Each 1.8 ohm resistor consumes 882 milliwatt.
  • Total power consumed by the resistors is 3528 milliwatt.
  • Total power consumed by the LEDs is 30240 milliwatt.
  • Total power consumed by the circuit is 33768 milliwatt.
  • Total current drawn by the circuit is 2800 milliampere.
  • The resistor values are calculated based on the common ±5% tolerance resistors.
  • Make sure to wire the LEDs in the correct direction as shown below.
  • Always leave some space for the resistors to breathe. They might get hot.
I have used 10 Watt LED’s.
When I first turn on the circuit all the LED’s Light they get hot around 40°c. Then one array dims and the other starts to flash and goes off completely.
Any advice will be greatly appreciated to resolve tis challenge.
Thank you
Rodney
 

Wendy

Joined Mar 24, 2008
23,415
You forgot little things. Like the power supply voltage. A schematic is also nice, people are always trying to put LEDs in direct parallel with each other, which is a really bad idea. Like transistors, no two LEDs are quite alike, so one hogs the current until it can not take any more, then the second blows after.

The chinese products get by with it because they match large lots to each other, which is not usual. They also don't care if an LED lasts beyond a month or two.
 

Thread Starter

RodneyB

Joined Apr 28, 2012
697
The Supply voltage is 12 Volts

I dont know how to post a schematic guidance please.

They are LED's from China
 

Thread Starter

RodneyB

Joined Apr 28, 2012
697
Just a note when I have a fan blowing on the circuit I dont have a problem. Switch the fan off and you can see smoke. After a few minutes the one array starts to flash. Then eventusally goes off. I currently have it running with the fan on again. No flashing
 

Wendy

Joined Mar 24, 2008
23,415
I translated your schematic into something a little more forum friendly (using M/S Paint and Print Scrn).



It is quite normal for 3W and 10W LEDs to get hot. Each LED is dissipating 3W, which can leave a nice blister on the finger. So heat sinking or some other means of cooling is required.

How sure of the .7A are you? Have you measured it?

You have probably done damage to the current LEDs. Don't worry about it too much, you just lost a couple of thousand hours out of 10's of thousands.

<edit>

Just read your last post, they may be toast. Cooling is critical, so is the current. With a resistance so low it is likely you are way out of the ball park on current. Most folks around here do not like LED calculators, as they have no common sense whatsoever.

Bill's Blog

LEDs, 555s, Flashers, and Light Chasers
 

Attachments

Wendy

Joined Mar 24, 2008
23,415
BTW, if you want help for the next time (I assume you will likely buy replacement LEDs) just ask, all the folk here are kind to beginners (one of my several jobs) and we like questions.

It would help if you could figure out how to attach pictures of your setup, you have figured out attachments. When you attach a picture you can put the address of the picture between the {img}IP address{/img} tags, only use brackets [] instead of {}.

I like helping people, but time has been an issue lately. I will try to check back on this thread later.

Don't feel bad, no one is born knowing this stuff.
 

Thread Starter

RodneyB

Joined Apr 28, 2012
697
Many thanks for the help. I made some very elementary mistakes when inputting the information into the LED calculator.
  • The battery voltage was actual 13.8 Volts so I used 14 Volts as the supply voltage.
  • I was driving the LED at the Imax as shown on the Data sheet.
  • I reduced the Imax to 350mA and the result was a 10R resistor at 5 watts. I used a 10R resistor at 10 Watts and it is working.
Yes it gets hot but no flashing or breaking down of LED’s I am going to back the PCB to a piece of Aluminium and then secure that to the lamp housing so effectively the whole lamp housing becomes a heat sink.
One thing that puzzles me is that the LED’s that I replaced still glow faintly I haven’t tested them individually but I was under the impression that when an led was blown, it stopped glowing.
Again my sincere thanks
Rodney
 

mcgyvr

Joined Oct 15, 2009
5,394
welcome to "thermal runaway" and more..
What you are doing is building a fire creator more than anything.. please stop and learn why before you continue..

resistors and constant voltage power supplies are a serious NO-NO with high power LED's (anything over 1W really) They must be driven with constant current power supplies.
Slight changes in the forward voltage of the LED's can cause large changes in the current per string.. large changes in that current can cause destruction of the LED's or worse..a fire that could burn your house down.. yes its that serious..
There is a reason they worked fine for a bit then went to hell.. changing forward voltage and thermal dissipation issues. aka thermal runaway more or less.

those led resistor calculators work fine for gum drop 20ma leds but are NOT intended for higher power leds.. again..do not use resistors.. and frankly avoid parallel strings too unless you have proper current mirroring protection in effect.
 

takao21203

Joined Apr 28, 2012
3,702
You can almost forget about to use power LEDs with batteries or linear supplies.

Unless you have a transformer that just bounces down right when you load it with some Amperes.

Even 0.3 to 0.5 volts can be critical.

You need at least a SMPS booster or buck. They are often somehow current limited, for instance by the coil size. Many also have an active current limit.

If you want a PERMANENT installation which will be unattended for weeks, months and years- use a proper LED driver with a good enclosure.

If you have a custom panel with many LEDs you may want to drive them with a specialized circuit.

These should by all means include an electronic transformer.

Resistors can only be used if the voltage absolutely does not change.

For a large power LED, a few toroid coils are sufficient here in one case to reduce the current to the right volume. 12V LED, but the 12V are just not right by some 100mV.

If LEDs turn dim or go out, you are totally overdriving them.

Each LED is different, do not trust the datasheet. You either must use a constant current driver, or you have to measure and test the individual assembly for heating, and current, and performance.

Voltages in the datasheet can be off by a volt or so- your LED will simply burn out.
 

Thread Starter

RodneyB

Joined Apr 28, 2012
697
I just read about thermal run away ...... mmmm .... scary. But now my curiosity has been sparked. How do I build a 1.5Amp current limiting circuit?

Where I live in Zimbabwe we have lots of power cuts and these lights were wanted for security as crime is on the increase and becoming violent. I want to see whats coming so I know what I have to protect myself from.
 

Wendy

Joined Mar 24, 2008
23,415
I will disagree with the statements about the power supplies. The real problem is head room. You can use a convention resistor just fine, but if you only allow 2 V a minor variation isn't anymore. You can measure the current by measuring the voltage across the resistor. The equation is fairly simple, I = V / R

V is the measure voltage across the resistor, R is the resistance of the resistor, and I is the measured current.

A linear regulator does not vary output voltages any more than a SMPS regulator, it just gets a lot hotter doing so. If the power supply voltage is stable then it does not matter where it comes from.

The OP has 3.6V X 3, or 10.8V. He also describes 13.2 V, voltage stability is critical with this. The head room described is 2.5V, which is a little close but should be enough to take up the slack.

If the power supply is 12VDC then the headroom is 1.2V, which is way too close.

The reason is the smaller the differences in voltage the more difference even a small change can make on the power supply. Resistor tolerances are also more critical for the same reasons.

The heat sink (the piece of aluminum you described) is critical, the circuit will not work long term without it. Heat is always the enemy in electronics, so any way you can get rid of it is important.
 

#12

Joined Nov 30, 2010
18,224
One more point for Bill's presentation...batch differences in the LEDs. The LEDs have a range of possible voltages from 3.4 volts to 3.8 volts. Quoting Bill, in this case, what would normally seem to be a minor variation isn't.
 

bertus

Joined Apr 5, 2008
22,270
Hello,

I always take the lowest given voltage for voltage drop of the led for calculating the series resistor.
That way the current can only get less if the voltage drop of the leds is higher.
This will be more safe then when you take the typical voltage.

Bertus
 

Thread Starter

RodneyB

Joined Apr 28, 2012
697
Where I live (zimbabwe) we are plagued by constant power cuts. As a result criminals are now using the dark to brake in and attack people and steal. I wanted these flood lights to be solar driven from a battery. So at least I can see what I am up against. Do you have a constant current circuit diagram that I can use. I read about thermal run away. Scary stuff
 

Thread Starter

RodneyB

Joined Apr 28, 2012
697
I reduced the current to half of that shown on the datasheet and it made a massive diffarence and lights still bright
 

takao21203

Joined Apr 28, 2012
3,702
2x 20W LEDs. They do not even turn warm.

60v is used they are in series. This has advantage, as each LED has some internal resistance.

At 10 LEDs you already get a comfortable headroom. Here are about 20 in series.

8 dollars voltage booster, needs a larger sink for the MOSFET, especially since it is patched to produce higher voltages.

Runs from a laptop adapter, 19.5V, 70W.

There is no such thing as a resistor. But I need a choke to limit the inrush current, or the laptop adapter can not startup, and the LEDs actually blink.

There is totally no possibility the LEDs can burn or take any damage.

A long time or professional setup of course must monitor the cooler temperature- for instance the fan could stop spinning one day.

The silicone cement works perfect actually, it is even fully heat resistant. Why screw LEDs?

Note this is not a long term installation. I am just starting to examine this technology. There are many cables floating around actually. One splash of water- everything is ruined.

Normally there should be only one cable, and the converter in a box close to the LEDs, the LEDs having an enclosure, diffusor or lens as well.

I am not a LED lamp factory...
 

Attachments

Top