Diode Regulation?

[twhitehouse]

Am I correct in the following?

A diode from Radioshack has a maximum Vf of 0.8, capable of handling 1A.

A Cree XP-E has a maximum Vf of 3.5, also capable of 1A.

If I put 4 diodes and 3 XP-Es in series, that would be a theoretical max of 0.8x4= 3.2Vf, plus 3.5x3=10.5Vf, total 13.7Vf.

Let's say they balance out a little under their maximums:

4 diodes at 0.6Vf, (2.4Vf) plus 3 XP-Es at 3.4Vf (10.2Vf) equals 12.6 Total Vf.

A fully charged project SLA battery is 12.6VDC (minus any surface charge on the plates due to charging).

So, if I put these 4 diodes and 3 XP-Es in series, it SHOULD theoretically be okay to direct drive them, as the total voltage drop across all elements is either equal to, or more than, what the SLA can push through.

According to my ASE A6 Electrical certification, this makes sense. However, LEDs aren't part of that test, and I want to be sure I'm not missing something particular about the way they work.

Assuming my calculations are accurate, and the diodes drop the voltage around .6 each, and the LEDs drop about 3.4 each, according to Cree's Datasheet, that would have the Crees drawing about 800mA, through the entire circuit (well under the 1A limit of the diodes).

BTW, the diodes in question are part# 1N4004.

I know, I know, I should be using a CC driver, specifically built for driving LEDs, and there's always the resistor option, which performs a similar function in this application as the diodes. I know this is not the most efficient setup, and I'm not taking into account voltage sag on the SLA when presented with a load. I just want to make sure my logic and theoretical numbers are sound.

Thanks!

 

[Eltehs]

So, if I put these 4 diodes and 3 XP-Es in series, it SHOULD theoretically be okay to direct drive them, as the total voltage drop across all elements is either equal to, or more than, what the SLA can push through.

You miss at least the following important thing: voltage drop on the diode is not constant. It depends of the diode junction temperature.
Check this my post for details:
http://www.candlepowerforums.com/vb/showpost.php?p=3544404&postcount=14

So, if you switch on your diodes, they would heat up and the voltage drop decreases. As soon, as voltage drop decreases, current (limiting only by the internal SLA battery in your application) significantly increases. See typical voltage-current graph below:

As soon, as current increases, LED will be heated faster and so on. So quickly LED current increases over maximum and :poof:

 

[Steve K]

plus, the Vf for any diode (led or plain rectifier) will vary slightly from part to part. Either a proper current regulator is needed, or allow some voltage headroom for adding a series resistor which will provide less variability in the current as the leds change temperature.

Practically speaking, the voltage that you would have dropped across the 1N4004's might as well be dropped across a resistor. The energy is being wasted anyway (i.e. not turned into light), so you might as well get some benefit from the stabilization of the current.

regards,
Steve K.

 

[MikeAusC]

All the interior lights in my car consist of strings of 3 LEDs plus a resistor in series - why - because it works, despite temperature and voltage variations.

Both diodes and LEDs have a negative temperature-voltage co-efficient, which as explained above will lead to VERY short LED life.