White 5mm led's with 2.27V Vf @20mA

[Amonra]

I had a few extra of these http://www.led1.de/shop/index.php?cName=led-5mm-ultra-bright-white-12000mcd-c-3_17 5mm led's so i decided to make an experiment.
I soldered a few of them in series and connected them directly to 240V (actually 238V) 50Hz AC mains power, as expected they flickered.
I soldered 70 in series with the assumption that they had an average Vf of about 3.3V @ 20mA, but when i checked the current drawn it was 70mA.
I then decided to add more led's in order to get to 20mA and ended up soldering a total of 100 led's in series with 20.7mA current drawn. So now the average Vf is about 2.38V which was strange.
I then decided to add a rectifier to the circuit as i did not like the flicker but now the output DC voltage and current from the rectifier dropped to 218V 16mA. I assumed that the rectifier is eating some voltage so i found that i had to reduce 4 led's to get back to 20.5mA and that means that the average Vf is now only about 2.27V !!!

This is very strange and i have the feeling that i am doing something wrong.
But if im not, this is awesome.

 

[space]

238VAC rectifed will give you ~335VDC. Divided by 96 LEDs gives you ~3,5Vf. Your 218V measurment must be wrong. 70 LEDs in series on 238 VAC give you 4,8Vf peak!!!

space

 

[jtr1962]

With 240VAC the peak voltage the LEDs are seeing is 338V less the losses in the rectifer (~1.5V). Remember that even with a rectifier the shape of the voltage waveform is still a sine wave, only with the negative portion rectifed so it's positive. Your voltmeter is only reading the average voltage which would indeed be around 216V to 218V (note that 240V is the RMS voltage and this IS NOT the same as the average voltage). The current through the LEDs is actually varying from a peak of whatever they would draw at a Vf of (338-1.5)/96 = 3.50V down to 0 when the sine wave is under roughly 230V. My guess is this would indeed average around 20 mA but the peak current will be very hard on the LEDs.

No white LED will produce any appreciable amount of light on 2.27VDC.

 

[Amonra]

i dont get it, why would 240V be 338V ?
I dont know much about AC, can someone enlighten me ?

 

[jtr1962]

i dont get it, why would 240V be 338V ?
I dont know much about AC, can someone enlighten me ?

http://en.wikipedia.org/wiki/Alternating_current

Go down to the part where it says "Mathematics of AC voltages". That should explain everything better than I can.

 

[IMSabbel]

i dont get it, why would 240V be 338V ?
I dont know much about AC, can someone enlighten me ?

One thing first: If you do not know the basics of electrics, please dont work on the high voltage side. For your own safety (directly connecting 70 leds in series might be electrocution waiting to happen if not done very right).

And about this question: 240V means 240V because it will create the same power on an ohm resistor as a DC of 240V.
As the form is of the current is a Sinus, to get the equivalent voltage you have to integrate over I*U over one cycle. This is an integral over sin^2(t), which is 1/Sqrt(2) times the peak value.

Thus a 240V AC line will go up to sqrt(2)*240V in the peaks, and with the steep Vf line of LEDs, all those 20mA can flow right at the peak time.