> Though the change in operating point due to heat actually makes things
> worse - the Vf of an LED decreases as it heats up, so the LED with the
> lowest initial Vf will take more of the current, heat up more, and
> decreases its Vf even more, making for a bit of a viscous cycle.

Yes. (A vicious one as well :-) ).

I see my text was ambiguous.

>> * I say somewhat stable as an LED will heat due to current drain which
>> will affect its operating point and so affect its relationships with
>> the other LEDs. Aslow dance can be expected but so slow and so small
>> in most cases as to not be overly noticeable or important.

In the context I meant "somewhat stable as opposed to stable" but I see
that that could be taken as eg "not as unstable as it otherwise might have
been" :-).

So, yes.
Effectively you y=3Dtend to get thermal runaway, saved mainly by the fact
that when lots of paralleling is done LEDs are usually small eg 5mmso not
too much power per case size. AND wgen people parallel large numbers they
are often somewhat thermally linked either on a pcb or within an enclosure.

In cheap junk lights with a common series resistor OR even no resistor at
all !!! you often see a much brighter LED. This will soon fade and/or die
and it's olace be taken by the next victim in the chain, but usually worse
as there are less LEs to share the same current amongst.



          Russell.






>
> Though the change in operating point due to heat actually makes things
> worse - the Vf of an LED decreases as it heats up, so the LED with the
> lowest initial Vf will take more of the current, heat up more, and
> decreases its Vf even more, making for a bit of a viscous cycle.
> Shouldn't make things hugely worse with paralleled LEDs, but it's
> something to bear in mind (and can be a big issue if attempting to
> drive LEDs from a fixed voltage with no series resistor).
>
> (I appreciate you know this, Russell, but it's as well to make others
> aware who might think that heating would decrease the current as is
> typical with a resistor).
>
> Chris
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.