30x-40x CREE/SSC in series directly into the wall with a rectifier

Would this work?

Do you mind 50 or 60 Hz flicker (100 or 120 if you use a full wave rectifier)? No flicker if you have a suitable capacitor at the output of the rectifier.

You would have to isolate the LEDs from their heatsinks or the heatsinks from the world to avoid shock hazards. I think CSA (Canadian Standards Association) requires the insulation to survive 2500V for a short duration.

You also might want some protection against transients so that a lightning surge or other power glitch doesn't blow out $200 worth of LEDs.

And you would want a suitable resistor to limit the current to a safe value. Or a current regulator.

Greg

It would be darn bright if it did!

No.

When you connect an LED across a fixed voltage source, it draws some amount of current which causes it to light and to heat up. Heating causes the current to increase rather dramatically. This causes the LED to heat up more, which increases the current more, and so forth. This phenomenon is called "thermal runaway" and often ends pretty spectacularly.

The mains provides a fixed voltage with a very large amount of available current. What you need to do is limit the maximum current you can get from the mains to something the LEDs will tolerate (about 1 amp if they have very good heat sinking). Better yet is to regulate the current. The simplest way would be to add a series resistor -- a conservative value would be about 1 ohm per LED, but you could probably get by with less. (At one ohm per LED, it would need to dissipate about one watt per LED if you're running them at 1 amp.) An even better current regulator is an incandescent bulb which you could put in series with the string.

You might also have to put a resistor and maybe capacitor across the LED string to force the reverse voltage to drop across the rectifier rather than the LEDs. If dropped across the LEDs, it's likely to cause unequal reverse voltages on the LEDs, which might destroy them.

I don't recommend connecting anything to the power mains unless you have a pretty clear idea of what you're doing. Enough energy is available there to cause some serious problems.

c_c

Curious_character said:

No.

When you connect an LED across a fixed voltage source, it draws some amount of current which causes it to light and to heat up. Heating causes the current to increase rather dramatically. This causes the LED to heat up more, which increases the current more, and so forth. This phenomenon is called "thermal runaway" and often ends pretty spectacularly.

The mains provides a fixed voltage with a very large amount of available current. What you need to do is limit the maximum current you can get from the mains to something the LEDs will tolerate (about 1 amp if they have very good heat sinking). Better yet is to regulate the current. The simplest way would be to add a series resistor -- a conservative value would be about 1 ohm per LED, but you could probably get by with less. (At one ohm per LED, it would need to dissipate about one watt per LED if you're running them at 1 amp.) An even better current regulator is an incandescent bulb which you could put in series with the string.

You might also have to put a resistor and maybe capacitor across the LED string to force the reverse voltage to drop across the rectifier rather than the LEDs. If dropped across the LEDs, it's likely to cause unequal reverse voltages on the LEDs, which might destroy them.

I don't recommend connecting anything to the power mains unless you have a pretty clear idea of what you're doing. Enough energy is available there to cause some serious problems.

c_c

Click to expand...

Don't worry, this was purely theoretical, and I enjoy my current life Thanks for the insight though... I was just curious if it was just that simple.

Not recommended, I one or two short out the rest will go BanG

TORCH_BOY said:

Not recommended, I one or two short out the rest will go BanG

Click to expand...

That loud, huh?

If i ever was about to hook up that much money in form of leds, i would spend the extra money on a real driver...

zeeexsixare said:

Would this work?

Click to expand...

i almost had this done, with a bridge rectafryer 55 1W Batwing luxeons, and a high voltage cap.
they calculated the dc voltage at ~170V

i got a 8' bar, and everything was going peachy, i had strung 55 of them suckers all bases Thermal epoxied to to the aluminum bar.

as the final test i would test sections of the thing, over and over, then increase the voltage, and keep testing.

it all went to heck in a handbasket, when i found out that the emitter bases werent nutreal :-( (neither is the seul, and the cree has to be "fixed" to get it isolated).

i was already going to have enough issues with 170V flying around :-(
i tore it all off, and wouldnt do it again till i can afford 300$ worth of STAR bases fully electrically isolated.

If you would like to donate i could finish it, i think it would be very efficient with full bridge rectification, and capacitor flattening.

zeeexsixare said:

directly into the wall

Click to expand...

What kind of wall and how are you mounting the LEDs. LEDs generate heat and if the heat cannot escape you have a fire.

I saw the following on the TV news last week:
A family had replaced all their light bulbs with compact flouresents. That evening they had a burning smell. Could not figure where it was comming from until the wife looked up at the pot lights buried in the ceiling. The junction where the tube entered the base of the lamps was burned black. 30 LEDs at 700mA is 75 watts, at 350mA 37.5 watts, much more than a typical CFL. Make sure you do your heat dissipation calculations as well as your electrical calculations.

Equipment powered from the mains require approval.
In Canada contact CSA
http://www.csa.ca/Default.asp?language=english
In the USA contact UL
http://www.ul.com/
In other countries find out who to contact by asking your electrical utility.

For reference reasons for CSA/UL
http://www.faqs.org/faqs/electrical-wiring/part1/section-8.html
http://www.faqs.org/faqs/electrical-wiring/part1/section-7.html

Well, I have a small 5mm LED night light that works as follows:

1) Large capacitor in series with the incoming AC -- this limits the current (capacitors limit AC current, similar to a resistor with DC)
2) Bridge rectifier and filter capacitor converts the AC into DC
3) A resistor in series with the LED for additional current limiting.
4) A transistor, controlled by a small "sense" current, switches the circuit to the LED off and on in order to maintain 25mA current.
5) A photoresistor is hooked up in parallel with the sense resistor, when the room is already lit up, the resistance in this part goes down, causing the current to the transistor to go up, and the circuit to shut off.

The same type of circuit could be scaled up to run a whole bunch of Cree/SSC in series. However, if you're just trying to flood light some big area, getting a good fluorescent fixture is a better idea, and you won't end up spending @ $500 to do it. I could get the same amount of lumens, with the same color temperature, and slightly better color rendering for around $30 -- the price of a fixture with an electronic ballast, and 2 T8 5000k bulbs. On top of that, you would be generating that light from less electricity than the LEDs.

Where LEDs would have an advantage IMO would be for track lighting, as they can be collimated into a spotlight and fluorescent can't (though for accent lighting, short T5 tubes with reflectors can look quite good) I would want a separate driver on each LED, and the ability to add/remove them at will, not a string of them in series.

If you are really interested in that project, I suggest you to make first a prototype using 5mm leds and lower current values.

I tryed to do something similar sometime ago, and I ended with a lot of useless 5mm leds because of the "thermal runaway" problem mentioned above.

I did a series of 8 "take-off" luxeons from a 27 voltage source on a heat sink. One LED popped. Replaced it, then it happened again. I didn't realize the chirping noise I was hearing was the LED frying. I though it was the voltage supply under load. The LED that ended up fried was due to thermal runaway. It was a fun experiment and didn't cost me anything but time.