LED string current - Absolute Max or a safer current ?

Other Hardware General Electronics
1

Greetings again, Gurus

I have made a hanging kitchen light for an extended table, with an aliminium H-shaped strip, holes drilled for LEDs ( thinking the aluminium will act as a heatsink, and a light reflector, and it looks good - even the wife likes it ).

Using White LEDs ( NSPW500DS http://docs-europe.electrocomponents.com/webdocs/0e76/0900766b80e762c9.pdf )

Absolute Max Fwd current ( IF ) = 30 mA
Forward Voltage ( VF ) = IF=20mA Typ. = 3.2V Max = 3.5V

LEDs are wired in series ( 3 per string ).

Obviously, I want to get the best lighting out of the LEDs.

So my question is, ( I am ) assuming the light emitted would be 'brighter' if I sized the resistors for IF = 26mA.

Is there a "safe" 'IF' to use ? ( wanting maximum lighting, but reasonable LED lifespan )
80% of max 'IF' ?
70% of max 'IF' ?

Once the 'safe' 'IF' has been decided, would this affect the calculations ?

Spec : Forward Voltage ( VF ) = IF=20mA Typ. = 3.2V Max = 3.5V
I imagine that the fact that they list the " IF=20mA " on this line in the specs, has some relevance on the "Typ. = 3.2V" ?

My initial testing had a supply of 11.21V and a 150 ohm resistor per string of 3 LEDs, which I calculate back to VF = 10.73 mA.

I have just changed my power supply from a wallwart to a central supply ( which has battery backup ) and now the supply is 13.27V, with 0.87V drop over the switch ( using tact switches and C106D Reverse Blocking Thyristors ), making the supply to the lighting = 12.4V.
With the existing 150 ohm resistors, the VF is now = 18.6 mA.

The increase in the lighting is quite noticable.

I am easily able to change the resistors, so was thinking that a 100 ohm would make 'IF' = 28 mA. ( too close to the max limit ? )

additional thought : or do I change to 2 x 220ohm resistors in parrallel, making an effective 110 ohms = 'IF' = 25.5 mA

2

20mA is the 'working' current. If you want more light you need to add more LEDs. Running them at 28mA is like driving your car around at 6000RPM - it works, but it's slowly killing your engine.

Me? I'd aim for 20mA (using 3.2V as the working voltage for the math).

Even better, use a proper constant-current circuit and set it to "20mA".

3

For a DIY project, I wouldn't be afraid to "push-it" a bit. I assume you have some spare LEDs... Worst case, one burns-up... You can replace it and increase the resistor values so it doesn't happen again.

If I was building those lights in quantity and selling them, I'd be more conservative.

It's like any other risk/reward "gambling situation"... If the odds of loosing are low and/or the cost of loosing is low, and the winning payoff is good, it's probably worth taking the risk!

LEDs are wired in series ( 3 per string ).

I don't know what typically happens when you over-current an LED. Most semiconductors tend to short-out when you over-voltage, over-current, or over-heat them. If one series LED shorts-out, that will put more current through the other two, which might burm them out out also.

4

Most LEDs show right in their spec sheet that running at high currents results in decreased life, due to overheating.
Keep it at 20mA.

5

Many Thanks to all for the replies.

Now I understand why.

20mA max, or 18.6 mA as it is now, will be the final current.

Long Live the LEDs

6

DVDdoug:
I don't know what typically happens when you over-current an LED. Most semiconductors tend to short-out when you over-voltage, over-current, or over-heat them. If one series LED shorts-out, that will put more current through the other two, which might burm them out out also.

All the damaged LEDs I've seen get dimmer before they die.

If they get dimmer when you damage them (if you put a new LED alongside you can see a difference). Getting dimmer when damaged kinda defeats the purpose of running them at higher current to get "extra brightness!!!", doesn't it?

7

fungus:
Getting dimmer when damaged kinda defeats the purpose of running them at higher current to get "extra brightness!!!", doesn't it?

Agreed. My dilemma was that I didn't know what was considered "higher current" and what was the best current for efficient lighting and getting long life from the LEDs.

My understanding now is that if IF=20mA Typ. = 3.2V then that is, I now assume, the recommended normal current for these LEDs. The 30mA listed in the Absolute Maximum spec is the "LED will DIE after this point" current.

It doesn't say that anything above 20mA is not ideal, but the advice from the gurus here would tend to indicate this.

Again I am grateful for the experienced and educated advice of the experts.

8

DaveO:
Many Thanks to all for the replies.

Now I understand why.

20mA max, or 18.6 mA as it is now, will be the final current.

Long Live the LEDs

When I want to light up a bunch of LEDs with no microcontroller involvement I use these:

(or ka2284 for sale | eBay - they both do exactly the same thing)

Yes, it's a VU meter but if you tie pin 7 to Vcc then all the outputs turn on. You get five constant-current sinks per chip, no need to do math for resistors. You can combine pins if you want more current (one chip can drive a 100mA LED). They fit vertically on a PCB so they don't take up much space (and need less soldering than resistors). If you want to switch/PWM them with an Arduino, you can... (pin 7 again), etc.

The only bummer is that they slightly under-power the LEDs (I guess they did this because 20mA is usually too bright to look at directly on modern LEDs).