Temperature coefficient diodes.

After mike did my head in again… (headache)

I decided to test the foward voltage drop of a regular diode, and zeners have the same effect, however reverse a zener to clamp down voltage and we have a voltage increase.

So this circuit (yet to build it… after opinions) is an attempt to supply a “regulated” current supply to the 1 watt led!

Any concerns?

cjdelphi: After mike did my head in again..... (headache)

I decided to test the foward voltage drop of a regular diode, and zeners have the same effect, however reverse a zener to clamp down voltage and we have a voltage increase.

So this circuit (yet to build it.... after opinions) is an attempt to supply a "regulated" current supply to the 1 watt led!

Any concerns?

Not sure whether there's a question here!

Forward biased diodes have a negative tempco, the voltage drops as the temperature increases as the difference between bandgap (1.2V for silicon) and forward voltage is roughly proportional to thermodynamic temperature (ie measured in kelvin).

Zener diode voltage tempco depends on the voltage rating and the type of "zener" (some are genuine zeners, some are avalanche-breakdown diodes). Middling voltage ones are alleged to have the smallest tempco, 6V or so, IIRC. Datasheets are always the place to check such details.

http://en.wikipedia.org/wiki/Diode#Shockley_diode_equation

I'm basically using the zener's reverse foward voltage drop to counter out diodes foward voltage drop.... so the voltage stays relatively stable.. pointless I know more curious in the way diodes are exploited like this even being used as a temp sensor (normal diode with a tiny current measuring the foward voltage drop)

The most problematic variation is the LED forward voltage itself! Temperature compensation for that would need thermal contact between LED die and compensating diode...

But the real issue is LEDs need constant current for constant light output - a constant current source will do the job without having to worry about temperature.

Any concerns?

A proper constant current or regulated current supply is going to require an active device... A transistor, MOSFET, or a chip with transistors/MOSFETs inside. And since linear designs tend to be very inefficient (and components tend to get hot), a switching design (with an inductor) is commonly used.

For regular low-power LEDs, we usually don't care about efficiency (and heat is not an issue) we can simply use a voltage-regulated supply and a current-limiting resistor. (The more voltage we drop across the resistor, the closer we come to a constant current supply, and the less efficient the circuit becomes.)

...If the voltage & current numbers on your schematic are correct, you've got less than 1/2 watt going to the LED (current x voltage) with more than 2W consumed by the other components.