On 04/01/15 10:10, Capitol wrote:
The Natural Philosopher wrote:
On 03/01/15 22:33, Adam Aglionby wrote:
On Saturday, January 3, 2015 9:11:17 PM UTC, Tim Watts wrote:
On 03/01/15 10:13, Tim Watts wrote:
On the subject - what *is* the best way to drive LEDs?

Seems to me that whilst convenient and in line with my earlier
comments
on standardisation, putting little 230V PSUs in every lamp that get
hot
and blow up is not the best way forward.

Does a 12V supply offer any advantages in terms of minimising on board
electronics? 12V SELV is at least standard.

24V is better , double voltage half the current and copper, but thats
laways the issue , low voltage dosen`t equal low current.


If an LED has a Vf (forward voltage drop) of x volts, is it considered
good form to put 12/x LEDs in series across the supply with no other
limiting circuitry?

Vf isn`t really fixed, it varies from batch to batch, fractionally but
enough.

Thing is a iny increase in Vf can have an expotentially large increase
in current, not so much of a pronlem at a nominal 20mA more of an
issue at 350mA


Or is there a really simple 2 pin current regulator on a chip
available?

2 pin will get very hot..

not of its a capacitor or an inductor.
frankly a choke and a full wave bridge and a smoothing cap is all you
need. Shove the LEDS in series.

Same as a fluorescent in principle, less the starter ********.
Fluorescent tube is just like an HV LED in general electrical
characteristics, once 'lit'.

Not very 'dimmable' mind you.




Old style 0.2" 20mA LEDs weren't that bothered, but I'm not au fait
with
high power Crees and the like.


for big LEDs constant current drive has lot to reccomend it over
constant voltage and resistors, little 20mA LEDs are fair game for all
sort sof things.

Anyone?




The thing about Vf is that has a negative temperature component in
the region of 2mV/C. So if you assume that there is a 50C operating
temperature variation, this changes Vf by 5%. This will produce massive
If variations if not fed from a fixed current supply. Any increase in If
increases running temperature, so you have perfect thermal runaway
conditions. LEDS do not like high temperatures, I can't remember all the
details now but the efficiency drops and the thermal losses increase,
hence the short life if you over run them.

that's why I suggested a series choke or cap: inherent current limit


--
Everything you read in newspapers is absolutely true, except for the
rare story of which you happen to have first-hand knowledge. €“ Erwin Knoll